diff --git a/packages/core/src/sync/index.ts b/packages/core/src/sync/index.ts
index a3fa90270f..a29c5aca5f 100644
--- a/packages/core/src/sync/index.ts
+++ b/packages/core/src/sync/index.ts
@@ -10,6 +10,7 @@ import { anthropic } from "./providers/anthropic.js";
 import { baseten } from "./providers/baseten.js";
 import { chutes } from "./providers/chutes.js";
 import { cloudflareWorkersAi } from "./providers/cloudflare-workers-ai.js";
+import { cortecs } from "./providers/cortecs.js";
 import { crossmodel } from "./providers/crossmodel.js";
 import { deepinfra } from "./providers/deepinfra.js";
 import { digitalocean } from "./providers/digitalocean.js";
@@ -116,6 +117,7 @@ export const providers: {
   baseten: SyncProvider<any>;
   chutes: SyncProvider<any>;
   "cloudflare-workers-ai": SyncProvider<any>;
+  cortecs: SyncProvider<any>;
   crossmodel: SyncProvider<any>;
   deepinfra: SyncProvider<any>;
   digitalocean: SyncProvider<any>;
@@ -144,6 +146,7 @@ export const providers: {
   baseten,
   chutes,
   "cloudflare-workers-ai": cloudflareWorkersAi,
+  cortecs,
   crossmodel,
   deepinfra,
   digitalocean,
@@ -183,7 +186,7 @@ export const groups = {
     "vercel",
   ],
   cloudflare: ["cloudflare-workers-ai"],
-  direct: ["ambient", "anthropic", "baseten", "chutes", "deepinfra", "digitalocean", "google", "hyper", "openai", "ovhcloud", "pioneer", "tinfoil", "venice", "wandb", "xai"],
+  direct: ["ambient", "anthropic", "baseten", "chutes", "cortecs", "deepinfra", "digitalocean", "google", "hyper", "openai", "ovhcloud", "pioneer", "tinfoil", "venice", "wandb", "xai"],
 } as const;
 
 type ProviderID = keyof typeof providers;
diff --git a/packages/core/src/sync/providers/cortecs.ts b/packages/core/src/sync/providers/cortecs.ts
new file mode 100644
index 0000000000..563c48d93a
--- /dev/null
+++ b/packages/core/src/sync/providers/cortecs.ts
@@ -0,0 +1,169 @@
+import { z } from "zod";
+
+import { describeModel } from "../../describe.js";
+import type { ExistingModel, SyncProvider, SyncedFullModel, SyncedModel } from "../index.js";
+import { factorBaseModel, resolveModelMetadataBaseModel } from "./openrouter.js";
+
+const API_ENDPOINT = "https://api.cortecs.ai/v1/models";
+const CANONICAL_BASE_MODEL_EXCEPTIONS = {
+  "claude-sonnet-4": "anthropic/claude-sonnet-4-0",
+} as const;
+// Cortecs publishes its default catalog prices in EUR per million tokens.
+// Exchange rate used by the existing Cortecs entries, as of 2026-07-30.
+const EUR_TO_USD = 1.114;
+
+const CortecsModality = z.enum(["text", "audio", "image", "video", "pdf"]);
+
+export const CortecsModel = z.object({
+  id: z.string().min(1),
+  created: z.number().int().nonnegative(),
+  description: z.string().optional(),
+  pricing: z.object({
+    currency: z.literal("EUR"),
+    input_token: z.number().nonnegative(),
+    output_token: z.number().nonnegative(),
+    cache_read_cost: z.number().nonnegative().optional(),
+    cache_write_cost: z.number().nonnegative().optional(),
+  }).passthrough(),
+  context_size: z.number().int().positive(),
+  input_modalities: z.array(CortecsModality).default(["text"]),
+  output_modalities: z.array(CortecsModality).default(["text"]),
+  supported_features: z.array(z.string()).default([]),
+}).passthrough();
+
+export const CortecsResponse = z.object({
+  object: z.literal("list"),
+  data: z.array(CortecsModel),
+}).passthrough();
+
+export type CortecsModel = z.infer<typeof CortecsModel>;
+
+export const cortecs = {
+  id: "cortecs",
+  name: "Cortecs",
+  modelsDir: "providers/cortecs/models",
+  deleteMissing: true,
+  async fetchModels() {
+    const response = await fetch(API_ENDPOINT);
+    if (!response.ok) {
+      throw new Error(`Cortecs models request failed: ${response.status} ${response.statusText}`);
+    }
+    return response.json();
+  },
+  parseModels(raw) {
+    return CortecsResponse.parse(raw).data;
+  },
+  translateModel(model, context) {
+    return {
+      id: model.id,
+      model: buildCortecsModel(model, context.existing(model.id), context.authored(model.id)),
+    };
+  },
+} satisfies SyncProvider<CortecsModel>;
+
+function dateFromTimestamp(timestamp: number) {
+  return new Date(timestamp * 1_000).toISOString().slice(0, 10);
+}
+
+function usd(value: number | undefined) {
+  if (value === undefined) return undefined;
+  return Math.round(value * EUR_TO_USD * 1_000) / 1_000;
+}
+
+export function buildCortecsModel(
+  model: CortecsModel,
+  existing: ExistingModel | undefined,
+  authored: ExistingModel | undefined,
+): SyncedModel {
+  const features = new Set(model.supported_features);
+  const input = model.input_modalities;
+  const output = model.output_modalities;
+  const canonical = existing?.base_model ?? resolveCortecsBaseModel(model.id);
+  const sourceReasoning = features.has("reasoning");
+  const reasoning = canonical === undefined ? sourceReasoning : existing?.reasoning ?? sourceReasoning;
+  const reasoningOptions = canonical === undefined
+    ? (sourceReasoning ? existing?.reasoning_options ?? [] : undefined)
+    : (existing?.reasoning === true ? existing.reasoning_options : undefined);
+  const limit = {
+    context: model.context_size,
+    input: existing?.limit?.input,
+    output: authored?.limit?.output ?? model.context_size,
+  };
+  const cost = {
+    input: usd(model.pricing.input_token),
+    output: usd(model.pricing.output_token),
+    cache_read: usd(model.pricing.cache_read_cost) ?? existing?.cost?.cache_read,
+    cache_write: usd(model.pricing.cache_write_cost) ?? existing?.cost?.cache_write,
+    reasoning: existing?.cost?.reasoning,
+    tiers: existing?.cost?.tiers,
+  };
+  if (canonical !== undefined) {
+    return factorBaseModel(canonical, {
+      description: existing?.description,
+      attachment: input.some((value) => value !== "text"),
+      reasoning: undefined,
+      reasoning_options: reasoningOptions,
+      temperature: existing?.temperature,
+      tool_call: features.has("tools"),
+      structured_output: features.has("json_mode"),
+      status: existing?.status,
+      interleaved: existing?.interleaved,
+      limit,
+      modalities: { input, output },
+      cost,
+    }, limit, existing?.base_model_omit);
+  }
+
+  const family = existing?.family;
+  return {
+    name: existing?.name ?? model.id,
+    description: existing?.description ?? model.description ?? describeModel({
+      id: model.id,
+      name: model.id,
+      family,
+      reasoning,
+      tool_call: features.has("tools"),
+      structured_output: features.has("json_mode"),
+      open_weights: existing?.open_weights ?? false,
+      limit,
+      modalities: { input, output },
+    }),
+    family,
+    release_date: existing?.release_date ?? dateFromTimestamp(model.created),
+    last_updated: existing?.last_updated ?? dateFromTimestamp(model.created),
+    attachment: input.some((value) => value !== "text"),
+    reasoning,
+    reasoning_options: reasoningOptions,
+    temperature: existing?.temperature ?? false,
+    tool_call: features.has("tools"),
+    structured_output: features.has("json_mode"),
+    knowledge: existing?.knowledge,
+    open_weights: existing?.open_weights ?? false,
+    status: existing?.status,
+    interleaved: existing?.interleaved,
+    cost,
+    limit,
+    modalities: { input, output },
+  } satisfies SyncedFullModel;
+}
+
+function resolveCortecsBaseModel(modelID: string) {
+  const exception = CANONICAL_BASE_MODEL_EXCEPTIONS[
+    modelID as keyof typeof CANONICAL_BASE_MODEL_EXCEPTIONS
+  ];
+  if (exception !== undefined) return resolveModelMetadataBaseModel(exception);
+
+  const trailingFamily = /^claude-(\d+)-(\d+)-(opus|sonnet|haiku)$/.exec(modelID);
+  if (trailingFamily !== null) {
+    const [, major, minor, family] = trailingFamily;
+    return resolveModelMetadataBaseModel(`anthropic/claude-${family}-${major}-${minor}`);
+  }
+
+  const compactFamily = /^claude-(opus|sonnet|haiku)(\d+)-(\d+)$/.exec(modelID);
+  if (compactFamily !== null) {
+    const [, family, major, minor] = compactFamily;
+    return resolveModelMetadataBaseModel(`anthropic/claude-${family}-${major}-${minor}`);
+  }
+
+  return resolveModelMetadataBaseModel(modelID);
+}
diff --git a/packages/core/test/sync.test.ts b/packages/core/test/sync.test.ts
index 10c7c62319..9e7d40d678 100644
--- a/packages/core/test/sync.test.ts
+++ b/packages/core/test/sync.test.ts
@@ -10,6 +10,7 @@ import {
   parseAnthropicPricing,
   type AnthropicModel,
 } from "../src/sync/providers/anthropic.js";
+import { buildCortecsModel, type CortecsModel } from "../src/sync/providers/cortecs.js";
 import {
   buildCrossModel,
   type CrossModelModel,
@@ -2138,6 +2139,27 @@ test("defaults new reasoning models to empty reasoning options", () => {
   });
 });
 
+test("preserves authored Cortecs reasoning options missing from the API", () => {
+  const model: CortecsModel = {
+    id: "deepseek-v4-flash-0731",
+    created: 1_775_088_000,
+    pricing: { currency: "EUR", input_token: 0.224, output_token: 0.269 },
+    context_size: 1_048_576,
+    input_modalities: ["text"],
+    output_modalities: ["text"],
+    supported_features: ["reasoning", "tools"],
+  };
+  const existing: ExistingModel = {
+    base_model: "deepseek/deepseek-v4-flash-0731",
+    reasoning: true,
+    reasoning_options: [{ type: "effort", values: ["low", "medium", "high"] }],
+  };
+
+  expect(buildCortecsModel(model, existing, existing)).toMatchObject({
+    reasoning_options: [{ type: "effort", values: ["low", "medium", "high"] }],
+  });
+});
+
 test("syncs OpenRouter reasoning efforts from model metadata", () => {
   const model = buildOpenRouterModel(openRouterModel({
     reasoning: {
diff --git a/providers/cortecs/models/apertus-70b.toml b/providers/cortecs/models/apertus-70b.toml
new file mode 100644
index 0000000000..84e15f30c3
--- /dev/null
+++ b/providers/cortecs/models/apertus-70b.toml
@@ -0,0 +1,23 @@
+name = "apertus-70b"
+description = "Apertus 70B is an open, multilingual language model designed for research, long-context reasoning, and sovereignty-focused AI systems."
+release_date = "2026-07-08"
+last_updated = "2026-07-08"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = false
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 1.393
+output = 2.228
+
+[limit]
+context = 65_536
+output = 65_536
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/claude-4-5-sonnet.toml b/providers/cortecs/models/claude-4-5-sonnet.toml
index 2998f5e300..dcbdccfe18 100644
--- a/providers/cortecs/models/claude-4-5-sonnet.toml
+++ b/providers/cortecs/models/claude-4-5-sonnet.toml
@@ -1,28 +1,23 @@
-name = "Claude 4.5 Sonnet"
+base_model = "anthropic/claude-sonnet-4-5"
 description = "Balanced Claude model for coding, analysis, agent workflows, and cost control"
-# Cortecs maps `reasoning_effort = low|medium|high` and
-# `thinking.budget_tokens >= 1024`; unsupported fields may be silently ignored.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-sonnet"
-release_date = "2025-09-29"
-last_updated = "2025-09-29"
-knowledge = "2025-07-31"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }, { type = "budget_tokens", min = 1_024 }]
-tool_call = true
-temperature = true
-open_weights = false
+structured_output = true
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 3.259
-output = 16.296
+input = 2.989
+output = 14.945
+cache_read = 0.326
+cache_write = 4.078
 
 [limit]
-context = 200_000
 output = 200_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-4-6-sonnet.toml b/providers/cortecs/models/claude-4-6-sonnet.toml
index 463e15b24e..a362064d1c 100644
--- a/providers/cortecs/models/claude-4-6-sonnet.toml
+++ b/providers/cortecs/models/claude-4-6-sonnet.toml
@@ -1,27 +1,23 @@
-name = "Claude Sonnet 4.6"
+base_model = "anthropic/claude-sonnet-4-6"
 description = "Balanced Claude model for coding, analysis, agent workflows, and cost control"
-# Cortecs maps `reasoning_effort = low|medium|high` and
-# `thinking.budget_tokens >= 1024`; unsupported fields may be silently ignored.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-sonnet"
-release_date = "2026-02-17"
-last_updated = "2026-03-13"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }, { type = "budget_tokens", min = 1_024 }]
-temperature = true
-tool_call = true
-knowledge = "2025-08-31"
-open_weights = false
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 3.59
-output = 17.92
+input = 3.196
+output = 15.94
+cache_read = 0.32
+cache_write = 3.999
 
 [limit]
-context = 1_000_000
 output = 1_000_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-haiku-4-5.toml b/providers/cortecs/models/claude-haiku-4-5.toml
index ae1150475b..b949ca1c12 100644
--- a/providers/cortecs/models/claude-haiku-4-5.toml
+++ b/providers/cortecs/models/claude-haiku-4-5.toml
@@ -1,27 +1,23 @@
-name = "Claude Haiku 4.5"
+base_model = "anthropic/claude-haiku-4-5"
 description = "Fast Claude model for responsive assistance, classification, and lightweight agents"
-# Cortecs maps `reasoning_effort = low|medium|high` and
-# `thinking.budget_tokens >= 1024`; unsupported fields may be silently ignored.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-haiku"
-release_date = "2025-10-15"
-last_updated = "2025-10-15"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }, { type = "budget_tokens", min = 1_024 }]
-temperature = true
-tool_call = true
-knowledge = "2025-02-28"
-open_weights = false
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 1.09
-output = 5.43
+input = 0.996
+output = 4.982
+cache_read = 0.099
+cache_write = 1.186
 
 [limit]
-context = 200_000
 output = 200_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-opus-5.toml b/providers/cortecs/models/claude-opus-5.toml
new file mode 100644
index 0000000000..30ea95e6a2
--- /dev/null
+++ b/providers/cortecs/models/claude-opus-5.toml
@@ -0,0 +1,22 @@
+base_model = "anthropic/claude-opus-5"
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
+
+[cost]
+input = 5.5
+output = 27.498
+cache_read = 0.55
+cache_write = 6.874
+
+[limit]
+output = 1_000_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-opus4-5.toml b/providers/cortecs/models/claude-opus4-5.toml
index 779f0209d4..e79b9bea47 100644
--- a/providers/cortecs/models/claude-opus4-5.toml
+++ b/providers/cortecs/models/claude-opus4-5.toml
@@ -1,27 +1,23 @@
-name = "Claude Opus 4.5"
+base_model = "anthropic/claude-opus-4-5"
 description = "Flagship Claude model for deep reasoning, coding, and long-horizon agents"
-# Cortecs maps `reasoning_effort = low|medium|high` and
-# `thinking.budget_tokens >= 1024`; unsupported fields may be silently ignored.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-opus"
-release_date = "2025-11-24"
-last_updated = "2025-11-24"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }, { type = "budget_tokens", min = 1_024 }]
-temperature = true
-tool_call = true
-knowledge = "2025-03-31"
-open_weights = false
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 5.98
-output = 29.89
+input = 5.313
+output = 26.568
+cache_read = 0.531
+cache_write = 6.645
 
 [limit]
-context = 200_000
 output = 200_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-opus4-6.toml b/providers/cortecs/models/claude-opus4-6.toml
index f7acd7d2c9..ef61043afb 100644
--- a/providers/cortecs/models/claude-opus4-6.toml
+++ b/providers/cortecs/models/claude-opus4-6.toml
@@ -1,27 +1,23 @@
-name = "Claude Opus 4.6"
+base_model = "anthropic/claude-opus-4-6"
 description = "Flagship Claude model for deep reasoning, coding, and long-horizon agents"
-# Cortecs maps `reasoning_effort = low|medium|high` and
-# `thinking.budget_tokens >= 1024`; unsupported fields may be silently ignored.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-opus"
-release_date = "2026-02-05"
-last_updated = "2026-03-13"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }, { type = "budget_tokens", min = 1_024 }]
-temperature = true
-tool_call = true
-knowledge = "2025-05-31"
-open_weights = false
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 5.98
-output = 29.89
+input = 5.313
+output = 26.561
+cache_read = 0.531
+cache_write = 6.645
 
 [limit]
-context = 1_000_000
 output = 1_000_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-opus4-7.toml b/providers/cortecs/models/claude-opus4-7.toml
index a607f98870..3208457cd6 100644
--- a/providers/cortecs/models/claude-opus4-7.toml
+++ b/providers/cortecs/models/claude-opus4-7.toml
@@ -1,29 +1,23 @@
-name = "Claude Opus 4.7"
+base_model = "anthropic/claude-opus-4-7"
 description = "Flagship Claude model for deep reasoning, coding, and long-horizon agents"
-# Cortecs maps `reasoning_effort = low|medium|high` to
-# `output_config.effort`; no explicit budget is exposed for this model.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "claude-opus"
-release_date = "2026-04-16"
-last_updated = "2026-04-16"
-attachment = true
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
-temperature = false
-tool_call = true
-knowledge = "2026-01-31"
-open_weights = false
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 5.60
-output = 27.99
-cache_read = 0.56
-cache_write = 6.99
+input = 5.437
+output = 27.186
+cache_read = 0.544
+cache_write = 6.797
 
 [limit]
-context = 1_000_000
-output = 128_000
+output = 1_000_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-opus4-8.toml b/providers/cortecs/models/claude-opus4-8.toml
index 2a78c94e5a..2f0993ec9c 100644
--- a/providers/cortecs/models/claude-opus4-8.toml
+++ b/providers/cortecs/models/claude-opus4-8.toml
@@ -1,11 +1,22 @@
 base_model = "anthropic/claude-opus-4-8"
-# Cortecs maps `reasoning_effort = low|medium|high` to
-# `output_config.effort`; no explicit budget is exposed for this model.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 5.64
-output = 28.198
-cache_read = 0.563
-cache_write = 7.049
+input = 5.437
+output = 27.186
+cache_read = 0.544
+cache_write = 6.797
+
+[limit]
+output = 1_000_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-sonnet-4.toml b/providers/cortecs/models/claude-sonnet-4.toml
index 05daba2439..a90031c2f6 100644
--- a/providers/cortecs/models/claude-sonnet-4.toml
+++ b/providers/cortecs/models/claude-sonnet-4.toml
@@ -1,24 +1,23 @@
-name = "Claude Sonnet 4"
+base_model = "anthropic/claude-sonnet-4-0"
 description = "Balanced Claude model for coding, analysis, agent workflows, and cost control"
-family = "claude-sonnet"
-release_date = "2025-05-22"
-last_updated = "2025-05-22"
-knowledge = "2025-03"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = false
+structured_output = true
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
 
 [cost]
-input = 3.307
-output = 16.536
+input = 2.898
+output = 14.493
+cache_read = 0.29
+cache_write = 3.624
 
 [limit]
-context = 200_000
-output = 64_000
+output = 200_000
 
 [modalities]
-input = ["text", "image", "pdf"]
-output = ["text"]
\ No newline at end of file
+input = ["text", "image"]
diff --git a/providers/cortecs/models/claude-sonnet-5.toml b/providers/cortecs/models/claude-sonnet-5.toml
new file mode 100644
index 0000000000..94f2c8cc3b
--- /dev/null
+++ b/providers/cortecs/models/claude-sonnet-5.toml
@@ -0,0 +1,22 @@
+base_model = "anthropic/claude-sonnet-5"
+structured_output = true
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[[reasoning_options]]
+type = "budget_tokens"
+min = 1_024
+
+[cost]
+input = 2.2
+output = 11
+cache_read = 0.219
+cache_write = 2.749
+
+[limit]
+output = 1_000_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/codestral-2508.toml b/providers/cortecs/models/codestral-2508.toml
index 248b6cf888..f9547b3dd4 100644
--- a/providers/cortecs/models/codestral-2508.toml
+++ b/providers/cortecs/models/codestral-2508.toml
@@ -3,17 +3,18 @@ description = "Mistral coding model for code completion, generation, and develop
 family = "mistral"
 release_date = "2025-07-30"
 last_updated = "2025-07-30"
-knowledge = "2025-03"
 attachment = false
 reasoning = false
-tool_call = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2025-03"
 open_weights = true
 
 [cost]
-input = 0.3
-output = 0.9
-cache_read = 0.03
+input = 0.334
+output = 1.003
+cache_read = 0.033
 
 [limit]
 context = 256_000
diff --git a/providers/cortecs/models/cosmos3-super-reasoner.toml b/providers/cortecs/models/cosmos3-super-reasoner.toml
new file mode 100644
index 0000000000..61bc01e31d
--- /dev/null
+++ b/providers/cortecs/models/cosmos3-super-reasoner.toml
@@ -0,0 +1,23 @@
+name = "cosmos3-super-reasoner"
+description = "Cosmos3 Super Reasoner is a high-capacity reasoning model designed for complex multi-agent tasks and advanced physical AI understanding."
+release_date = "2026-06-02"
+last_updated = "2026-06-02"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.099
+output = 0.296
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/deepseek-r1-0528.toml b/providers/cortecs/models/deepseek-r1-0528.toml
index acf9189393..6922ad9d38 100644
--- a/providers/cortecs/models/deepseek-r1-0528.toml
+++ b/providers/cortecs/models/deepseek-r1-0528.toml
@@ -3,17 +3,22 @@ description = "DeepSeek reasoning model for multi-step analysis, math, coding, a
 family = "deepseek-thinking"
 release_date = "2025-05-28"
 last_updated = "2025-05-28"
-knowledge = "2024-07"
 attachment = false
 reasoning = true
-reasoning_options = []
-tool_call = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2024-07"
 open_weights = true
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
 [cost]
-input = 0.585
-output = 2.307
+input = 0.652
+output = 2.57
+cache_read = 0.163
 
 [limit]
 context = 164_000
diff --git a/providers/cortecs/models/deepseek-v3-0324.toml b/providers/cortecs/models/deepseek-v3-0324.toml
deleted file mode 100644
index 814edea1a8..0000000000
--- a/providers/cortecs/models/deepseek-v3-0324.toml
+++ /dev/null
@@ -1,24 +0,0 @@
-name = "DeepSeek V3 0324"
-description = "DeepSeek chat model for instruction following, coding, and analysis"
-family = "deepseek"
-release_date = "2025-03-24"
-last_updated = "2025-03-24"
-knowledge = "2024-07"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
-
-[cost]
-input = 0.551
-output = 1.654
-
-[limit]
-context = 128_000
-output = 128_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
\ No newline at end of file
diff --git a/providers/cortecs/models/deepseek-v3.2.toml b/providers/cortecs/models/deepseek-v3.2.toml
index 3041e03774..a7a442fb32 100644
--- a/providers/cortecs/models/deepseek-v3.2.toml
+++ b/providers/cortecs/models/deepseek-v3.2.toml
@@ -3,17 +3,22 @@ description = "DeepSeek chat model for instruction following, coding, and analys
 family = "deepseek"
 release_date = "2025-12-01"
 last_updated = "2025-12-01"
-knowledge = "2024-07"
 attachment = false
 reasoning = true
-reasoning_options = []
-tool_call = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2024-07"
 open_weights = true
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
 [cost]
-input = 0.266
-output = 0.444
+input = 0.296
+output = 0.495
+cache_read = 0.075
 
 [limit]
 context = 163_840
diff --git a/providers/cortecs/models/deepseek-v4-flash.toml b/providers/cortecs/models/deepseek-v4-flash.toml
index 0737271382..b0369575e2 100644
--- a/providers/cortecs/models/deepseek-v4-flash.toml
+++ b/providers/cortecs/models/deepseek-v4-flash.toml
@@ -1,16 +1,18 @@
 base_model = "deepseek/deepseek-v4-flash"
-# Cortecs Chat maps `reasoning_effort = low|medium|high`; no budget is exposed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
 base_model_omit = ["structured_output"]
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
 
 [interleaved]
 field = "reasoning_content"
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
 [cost]
-input = 0.133
-output = 0.266
-cache_read = 0.0028
+input = 0.148
+output = 0.296
+cache_read = 0.037
 
 [limit]
 context = 1_048_576
+output = 1_048_576
diff --git a/providers/cortecs/models/deepseek-v4-pro.toml b/providers/cortecs/models/deepseek-v4-pro.toml
index 74f1cba5de..7fcc764658 100644
--- a/providers/cortecs/models/deepseek-v4-pro.toml
+++ b/providers/cortecs/models/deepseek-v4-pro.toml
@@ -1,16 +1,18 @@
 base_model = "deepseek/deepseek-v4-pro"
-# Cortecs Chat maps `reasoning_effort = low|medium|high`; no budget is exposed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
 base_model_omit = ["structured_output"]
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
 
 [interleaved]
 field = "reasoning_content"
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
 [cost]
-input = 1.553
-output = 3.106
-cache_read = 0.003625
+input = 1.73
+output = 3.46
+cache_read = 0.432
 
 [limit]
 context = 1_048_576
+output = 1_048_576
diff --git a/providers/cortecs/models/devstral-2512.toml b/providers/cortecs/models/devstral-2512.toml
index aa83b4edb1..833006d927 100644
--- a/providers/cortecs/models/devstral-2512.toml
+++ b/providers/cortecs/models/devstral-2512.toml
@@ -1,23 +1,12 @@
-name = "Devstral 2 2512"
+base_model = "mistral/devstral-2512"
 description = "Mistral coding agent model for repository tasks and software engineering workflows"
-release_date = "2025-12-09"
-last_updated = "2025-12-09"
-knowledge = "2025-12"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
+structured_output = true
 
 [cost]
-input = 0
-output = 0
+input = 0.446
+output = 2.228
+cache_read = 0.045
 
 [limit]
 context = 262_000
 output = 262_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
\ No newline at end of file
diff --git a/providers/cortecs/models/gemini-2.5-flash.toml b/providers/cortecs/models/gemini-2.5-flash.toml
new file mode 100644
index 0000000000..17f8d09329
--- /dev/null
+++ b/providers/cortecs/models/gemini-2.5-flash.toml
@@ -0,0 +1,14 @@
+base_model = "google/gemini-2.5-flash"
+reasoning_options = []
+
+[cost]
+input = 0.299
+output = 2.491
+cache_read = 0.029
+cache_write = 0.097
+
+[limit]
+output = 1_048_576
+
+[modalities]
+input = ["text", "image", "audio"]
diff --git a/providers/cortecs/models/gemini-2.5-pro.toml b/providers/cortecs/models/gemini-2.5-pro.toml
index 70c17b2180..b5fcdcaa34 100644
--- a/providers/cortecs/models/gemini-2.5-pro.toml
+++ b/providers/cortecs/models/gemini-2.5-pro.toml
@@ -1,24 +1,15 @@
-name = "Gemini 2.5 Pro"
+base_model = "google/gemini-2.5-pro"
 description = "Advanced Gemini model for complex reasoning, coding, and multimodal analysis"
-family = "gemini-pro"
-release_date = "2025-03-20"
-last_updated = "2025-06-17"
-knowledge = "2025-01"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = false
-
+reasoning_options = []
 
 [cost]
-input = 1.654
-output = 11.024
+input = 1.495
+output = 9.964
+cache_read = 0.242
+cache_write = 0.434
 
 [limit]
-context = 1_048_576
 output = 65_535
 
 [modalities]
-input = ["text", "image"]
-output = ["text"]
\ No newline at end of file
+input = ["text", "image", "audio"]
diff --git a/providers/cortecs/models/gemini-3.1-flash-lite.toml b/providers/cortecs/models/gemini-3.1-flash-lite.toml
new file mode 100644
index 0000000000..e7d859c978
--- /dev/null
+++ b/providers/cortecs/models/gemini-3.1-flash-lite.toml
@@ -0,0 +1,14 @@
+base_model = "google/gemini-3.1-flash-lite"
+reasoning_options = []
+
+[cost]
+input = 0.272
+output = 1.631
+cache_read = 0.025
+cache_write = 0.082
+
+[limit]
+output = 1_048_576
+
+[modalities]
+input = ["text", "image", "audio"]
diff --git a/providers/cortecs/models/gemini-3.5-flash.toml b/providers/cortecs/models/gemini-3.5-flash.toml
new file mode 100644
index 0000000000..6fe5c06dc7
--- /dev/null
+++ b/providers/cortecs/models/gemini-3.5-flash.toml
@@ -0,0 +1,14 @@
+base_model = "google/gemini-3.5-flash"
+reasoning_options = []
+
+[cost]
+input = 1.483
+output = 8.898
+cache_read = 0.148
+cache_write = 0.988
+
+[limit]
+output = 1_048_576
+
+[modalities]
+input = ["text", "image", "audio"]
diff --git a/providers/cortecs/models/gemma-3-27b-it.toml b/providers/cortecs/models/gemma-3-27b-it.toml
new file mode 100644
index 0000000000..29c974f0a5
--- /dev/null
+++ b/providers/cortecs/models/gemma-3-27b-it.toml
@@ -0,0 +1,23 @@
+name = "gemma-3-27b-it"
+description = "Gemma 3 is a family of lightweight, multimodal models from Google, supporting text and image inputs, multilingual capabilities, and a 131K context window."
+release_date = "2025-03-12"
+last_updated = "2025-03-12"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = false
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.099
+output = 0.299
+
+[limit]
+context = 131_000
+output = 131_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/gemma-4-26b-a4b-it.toml b/providers/cortecs/models/gemma-4-26b-a4b-it.toml
new file mode 100644
index 0000000000..dd9940d78d
--- /dev/null
+++ b/providers/cortecs/models/gemma-4-26b-a4b-it.toml
@@ -0,0 +1,10 @@
+base_model = "google/gemma-4-26b-a4b-it"
+reasoning_options = []
+
+[cost]
+input = 0.111
+output = 0.557
+
+[limit]
+context = 262_000
+output = 262_000
diff --git a/providers/cortecs/models/gemma-4-31b-it.toml b/providers/cortecs/models/gemma-4-31b-it.toml
new file mode 100644
index 0000000000..e73b09d7e6
--- /dev/null
+++ b/providers/cortecs/models/gemma-4-31b-it.toml
@@ -0,0 +1,10 @@
+base_model = "google/gemma-4-31b-it"
+reasoning_options = []
+
+[cost]
+input = 0.223
+output = 0.39
+
+[limit]
+context = 262_000
+output = 262_000
diff --git a/providers/cortecs/models/glm-4.5-air.toml b/providers/cortecs/models/glm-4.5-air.toml
deleted file mode 100644
index 43160d878d..0000000000
--- a/providers/cortecs/models/glm-4.5-air.toml
+++ /dev/null
@@ -1,24 +0,0 @@
-name = "GLM 4.5 Air"
-description = "Efficient GLM model for fast reasoning, coding, and agent workflows"
-family = "glm-air"
-release_date = "2025-08-01"
-last_updated = "2025-08-01"
-attachment = false
-reasoning = true
-reasoning_options = []
-temperature = true
-tool_call = true
-knowledge = "2025-04"
-open_weights = true
-
-[cost]
-input = 0.22
-output = 1.34
-
-[limit]
-context = 131_072
-output = 131_072
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/glm-4.5.toml b/providers/cortecs/models/glm-4.5.toml
deleted file mode 100644
index afb4f2a0fa..0000000000
--- a/providers/cortecs/models/glm-4.5.toml
+++ /dev/null
@@ -1,27 +0,0 @@
-name = "GLM 4.5"
-description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering"
-family = "glm"
-release_date = "2025-07-29"
-last_updated = "2025-07-29"
-attachment = false
-reasoning = true
-reasoning_options = []
-temperature = true
-tool_call = true
-knowledge = "2025-04"
-open_weights = true
-
-[cost]
-input = 0.67
-output = 2.46
-
-[limit]
-context = 131_072
-output = 131_072
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/glm-4.7-flash.toml b/providers/cortecs/models/glm-4.7-flash.toml
index cd9e7598cd..f739f88bdf 100644
--- a/providers/cortecs/models/glm-4.7-flash.toml
+++ b/providers/cortecs/models/glm-4.7-flash.toml
@@ -1,27 +1,15 @@
-name = "GLM-4.7-Flash"
+base_model = "zhipuai/glm-4.7-flash"
 description = "Efficient GLM model for fast reasoning, coding, and agent workflows"
-family = "glm"
-release_date = "2025-08-08"
-last_updated = "2025-08-08"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-tool_call = true
-knowledge = "2025-04"
-open_weights = true
 
 [interleaved]
 field = "reasoning_content"
 
 [cost]
-input = 0.09
-output = 0.53
+input = 0.08
+output = 0.478
 
 [limit]
 context = 203_000
 output = 203_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/glm-4.7.toml b/providers/cortecs/models/glm-4.7.toml
index 44071c68a8..18cd487e49 100644
--- a/providers/cortecs/models/glm-4.7.toml
+++ b/providers/cortecs/models/glm-4.7.toml
@@ -1,27 +1,15 @@
-name = "GLM 4.7"
+base_model = "zhipuai/glm-4.7"
 description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering"
-family = "glm"
-release_date = "2025-12-22"
-last_updated = "2025-12-22"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-tool_call = true
-knowledge = "2025-04"
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
 
 [cost]
-input = 0.45
-output = 2.23
+input = 0.78
+output = 2.785
 
 [limit]
-context = 198_000
+context = 202_752
 output = 198_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/glm-5-turbo.toml b/providers/cortecs/models/glm-5-turbo.toml
index 784b143f53..19c915b56f 100644
--- a/providers/cortecs/models/glm-5-turbo.toml
+++ b/providers/cortecs/models/glm-5-turbo.toml
@@ -2,7 +2,11 @@ base_model = "zhipuai/glm-5-turbo"
 reasoning_options = []
 
 [cost]
-input = 1.235
-output = 4.118
-cache_read = 0.308
+input = 1.186
+output = 3.955
+cache_read = 0.296
 cache_write = 1.544
+
+[limit]
+context = 202_752
+output = 202_752
diff --git a/providers/cortecs/models/glm-5.1.toml b/providers/cortecs/models/glm-5.1.toml
index 8aab7c54b8..98fcb2cd4d 100644
--- a/providers/cortecs/models/glm-5.1.toml
+++ b/providers/cortecs/models/glm-5.1.toml
@@ -1,28 +1,15 @@
-name = "GLM-5.1"
+base_model = "zhipuai/glm-5.1"
 description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering"
-family = "glm"
-release_date = "2026-04-14"
-last_updated = "2026-04-14"
-attachment = false
-reasoning = true
 reasoning_options = []
-structured_output = true
-temperature = true
-tool_call = true
-open_weights = true
 
 [interleaved]
 field = "reasoning_content"
 
 [cost]
-input = 1.31
-output = 4.10
-cache_read = 0.24
+input = 1.384
+output = 4.348
+cache_read = 0.346
 
 [limit]
-context = 204_800
-output = 131_072
-
-[modalities]
-input = ["text"]
-output = ["text"]
+context = 202_752
+output = 202_752
diff --git a/providers/cortecs/models/glm-5.2.toml b/providers/cortecs/models/glm-5.2.toml
index 305e7fbb98..48706497a1 100644
--- a/providers/cortecs/models/glm-5.2.toml
+++ b/providers/cortecs/models/glm-5.2.toml
@@ -1,12 +1,17 @@
 base_model = "zhipuai/glm-5.2"
-# Cortecs Chat maps `reasoning_effort = high|max`; other efforts are not listed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-reasoning_options = [{ type = "effort", values = ["high", "max"]}]
 
 [interleaved]
 field = "reasoning_content"
 
+[[reasoning_options]]
+type = "effort"
+values = ["high", "max"]
+
 [cost]
-input = 1.44
-output = 4.53
-cache_read = 0.39
+input = 1.2
+output = 4.2
+cache_read = 0.26
+
+[limit]
+context = 1_048_576
+output = 1_048_576
diff --git a/providers/cortecs/models/glm-5.toml b/providers/cortecs/models/glm-5.toml
index fdc8b8093a..bad6f314b7 100644
--- a/providers/cortecs/models/glm-5.toml
+++ b/providers/cortecs/models/glm-5.toml
@@ -1,26 +1,16 @@
-name = "GLM 5"
+base_model = "zhipuai/glm-5"
 description = "Flagship GLM model for hybrid reasoning, coding, and agentic engineering"
-family = "glm"
-release_date = "2026-02-11"
-last_updated = "2026-02-11"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
 
 [cost]
-input = 1.08
-output = 3.44
+input = 0.988
+output = 3.164
+cache_read = 0.247
 
 [limit]
 context = 202_752
 output = 202_752
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/glm-5v-turbo.toml b/providers/cortecs/models/glm-5v-turbo.toml
index 0dfcc2ac0d..033265446b 100644
--- a/providers/cortecs/models/glm-5v-turbo.toml
+++ b/providers/cortecs/models/glm-5v-turbo.toml
@@ -1,8 +1,16 @@
 base_model = "zhipuai/glm-5v-turbo"
+structured_output = true
 reasoning_options = []
 
 [cost]
-input = 1.235
-output = 4.118
-cache_read = 0.308
+input = 1.186
+output = 3.955
+cache_read = 0.296
 cache_write = 1.544
+
+[limit]
+context = 202_752
+output = 202_752
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-4.1-mini.toml b/providers/cortecs/models/gpt-4.1-mini.toml
new file mode 100644
index 0000000000..9c8840d630
--- /dev/null
+++ b/providers/cortecs/models/gpt-4.1-mini.toml
@@ -0,0 +1,13 @@
+base_model = "openai/gpt-4.1-mini"
+structured_output = false
+
+[cost]
+input = 0.434
+output = 1.704
+cache_read = 0.134
+
+[limit]
+output = 1_047_576
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-4.1-nano.toml b/providers/cortecs/models/gpt-4.1-nano.toml
new file mode 100644
index 0000000000..f2f678a0ac
--- /dev/null
+++ b/providers/cortecs/models/gpt-4.1-nano.toml
@@ -0,0 +1,10 @@
+base_model = "openai/gpt-4.1-nano"
+structured_output = false
+
+[cost]
+input = 0.111
+output = 0.434
+cache_read = 0.056
+
+[limit]
+output = 1_047_576
diff --git a/providers/cortecs/models/gpt-4.1.toml b/providers/cortecs/models/gpt-4.1.toml
index fd3c95da9c..f04a5a4b16 100644
--- a/providers/cortecs/models/gpt-4.1.toml
+++ b/providers/cortecs/models/gpt-4.1.toml
@@ -1,24 +1,14 @@
-name = "GPT 4.1"
+base_model = "openai/gpt-4.1"
 description = "GPT model for general reasoning, writing, coding, and tool-assisted tasks"
-family = "gpt"
-release_date = "2025-04-14"
-last_updated = "2025-04-14"
-knowledge = "2024-06"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = false
-
+structured_output = false
 
 [cost]
-input = 2.354
-output = 9.417
+input = 2.192
+output = 8.769
+cache_read = 0.546
 
 [limit]
-context = 1_047_576
-output = 32_768
+output = 1_047_576
 
 [modalities]
 input = ["text", "image"]
-output = ["text"]
\ No newline at end of file
diff --git a/providers/cortecs/models/gpt-4o-mini.toml b/providers/cortecs/models/gpt-4o-mini.toml
new file mode 100644
index 0000000000..7792656c8f
--- /dev/null
+++ b/providers/cortecs/models/gpt-4o-mini.toml
@@ -0,0 +1,13 @@
+base_model = "openai/gpt-4o-mini"
+structured_output = false
+
+[cost]
+input = 0.159
+output = 0.638
+cache_read = 0.081
+
+[limit]
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-4o.toml b/providers/cortecs/models/gpt-4o.toml
new file mode 100644
index 0000000000..eeea5e7f41
--- /dev/null
+++ b/providers/cortecs/models/gpt-4o.toml
@@ -0,0 +1,13 @@
+base_model = "openai/gpt-4o"
+structured_output = false
+
+[cost]
+input = 2.659
+output = 10.635
+cache_read = 1.33
+
+[limit]
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-5-mini.toml b/providers/cortecs/models/gpt-5-mini.toml
new file mode 100644
index 0000000000..e23c43c914
--- /dev/null
+++ b/providers/cortecs/models/gpt-5-mini.toml
@@ -0,0 +1,14 @@
+base_model = "openai/gpt-5-mini"
+structured_output = false
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 0.279
+output = 2.192
+cache_read = 0.056
+
+[limit]
+output = 400_000
diff --git a/providers/cortecs/models/gpt-5-nano.toml b/providers/cortecs/models/gpt-5-nano.toml
new file mode 100644
index 0000000000..b2d796c604
--- /dev/null
+++ b/providers/cortecs/models/gpt-5-nano.toml
@@ -0,0 +1,14 @@
+base_model = "openai/gpt-5-nano"
+structured_output = false
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 0.06
+output = 0.439
+cache_read = 0.019
+
+[limit]
+output = 400_000
diff --git a/providers/cortecs/models/gpt-5.1.toml b/providers/cortecs/models/gpt-5.1.toml
new file mode 100644
index 0000000000..c03a65e1b7
--- /dev/null
+++ b/providers/cortecs/models/gpt-5.1.toml
@@ -0,0 +1,14 @@
+base_model = "openai/gpt-5.1"
+structured_output = false
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 1.375
+output = 10.96
+cache_read = 0.156
+
+[limit]
+output = 400_000
diff --git a/providers/cortecs/models/gpt-5.4.toml b/providers/cortecs/models/gpt-5.4.toml
index 9dbc108e74..662683f69c 100644
--- a/providers/cortecs/models/gpt-5.4.toml
+++ b/providers/cortecs/models/gpt-5.4.toml
@@ -1,14 +1,17 @@
 base_model = "openai/gpt-5.4"
-# Cortecs Chat maps `reasoning_effort = low|medium|high`; no budget is exposed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
 base_model_omit = ["limit.input"]
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
 
 [cost]
-input = 3
-output = 16.13
-cache_read = 0.25
+input = 2.898
+output = 15.453
+cache_read = 0.242
 
 [limit]
-context = 1_050_000
-output = 128_000
+output = 1_050_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-5.6-luna.toml b/providers/cortecs/models/gpt-5.6-luna.toml
new file mode 100644
index 0000000000..66c10bec00
--- /dev/null
+++ b/providers/cortecs/models/gpt-5.6-luna.toml
@@ -0,0 +1,16 @@
+base_model = "openai/gpt-5.6-luna"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 1.1
+output = 6.599
+cache_read = 0.11
+
+[limit]
+output = 1_050_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-5.6-sol.toml b/providers/cortecs/models/gpt-5.6-sol.toml
new file mode 100644
index 0000000000..2cde59b712
--- /dev/null
+++ b/providers/cortecs/models/gpt-5.6-sol.toml
@@ -0,0 +1,16 @@
+base_model = "openai/gpt-5.6-sol"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 5.5
+output = 32.998
+cache_read = 0.55
+
+[limit]
+output = 1_050_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-5.6-terra.toml b/providers/cortecs/models/gpt-5.6-terra.toml
new file mode 100644
index 0000000000..27fded7324
--- /dev/null
+++ b/providers/cortecs/models/gpt-5.6-terra.toml
@@ -0,0 +1,16 @@
+base_model = "openai/gpt-5.6-terra"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 2.749
+output = 16.498
+cache_read = 0.275
+
+[limit]
+output = 1_050_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/gpt-5.toml b/providers/cortecs/models/gpt-5.toml
new file mode 100644
index 0000000000..32fd3f83da
--- /dev/null
+++ b/providers/cortecs/models/gpt-5.toml
@@ -0,0 +1,14 @@
+base_model = "openai/gpt-5"
+structured_output = false
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 1.375
+output = 10.96
+cache_read = 0.156
+
+[limit]
+output = 400_000
diff --git a/providers/cortecs/models/gpt-oss-120b.toml b/providers/cortecs/models/gpt-oss-120b.toml
index 027de59f60..7391e05e7e 100644
--- a/providers/cortecs/models/gpt-oss-120b.toml
+++ b/providers/cortecs/models/gpt-oss-120b.toml
@@ -1,27 +1,15 @@
-name = "GPT Oss 120b"
+base_model = "openai/gpt-oss-120b"
 description = "Open-weight GPT model for self-hosted reasoning and instruction-following workloads"
-# Cortecs Chat maps `reasoning_effort = low|medium|high`; no budget is exposed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "gpt-oss"
-release_date = "2025-08-05"
-last_updated = "2025-08-05"
-knowledge = "2024-01"
-attachment = false
-reasoning = true
-reasoning_options = [{ type = "effort", values = ["low", "medium", "high"] }]
-tool_call = true
-temperature = true
-open_weights = true
 
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
 
 [cost]
-input = 0
-output = 0
+input = 0.089
+output = 0.446
+cache_read = 0.01
 
 [limit]
-context = 128_000
+context = 131_000
 output = 128_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/gpt-oss-20b.toml b/providers/cortecs/models/gpt-oss-20b.toml
new file mode 100644
index 0000000000..4fb1f55cb7
--- /dev/null
+++ b/providers/cortecs/models/gpt-oss-20b.toml
@@ -0,0 +1,13 @@
+base_model = "openai/gpt-oss-20b"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 0.045
+output = 0.167
+
+[limit]
+context = 131_000
+output = 131_000
diff --git a/providers/cortecs/models/gpt-oss-safeguard-120b.toml b/providers/cortecs/models/gpt-oss-safeguard-120b.toml
new file mode 100644
index 0000000000..8789fbc555
--- /dev/null
+++ b/providers/cortecs/models/gpt-oss-safeguard-120b.toml
@@ -0,0 +1,13 @@
+base_model = "openai/gpt-oss-safeguard-120b"
+
+[[reasoning_options]]
+type = "effort"
+values = ["low", "medium", "high"]
+
+[cost]
+input = 0.179
+output = 0.697
+
+[limit]
+context = 128_000
+output = 128_000
diff --git a/providers/cortecs/models/hermes-4-405b.toml b/providers/cortecs/models/hermes-4-405b.toml
new file mode 100644
index 0000000000..bca06b6647
--- /dev/null
+++ b/providers/cortecs/models/hermes-4-405b.toml
@@ -0,0 +1,22 @@
+name = "hermes-4-405b"
+description = "Hermes 4 405B is a frontier hybrid-mode reasoning model built on Llama 3.1, optimized for advanced logic, math, coding, and structured output generation."
+release_date = "2024-08-13"
+last_updated = "2024-08-13"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.996
+output = 2.989
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/hermes-4-70b.toml b/providers/cortecs/models/hermes-4-70b.toml
index 4512dd0b0d..e640f0f67d 100644
--- a/providers/cortecs/models/hermes-4-70b.toml
+++ b/providers/cortecs/models/hermes-4-70b.toml
@@ -2,17 +2,17 @@ name = "Hermes 4 70B"
 description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use"
 release_date = "2025-08-26"
 last_updated = "2025-08-26"
-knowledge = "2023-12"
 attachment = false
-reasoning = true
-reasoning_options = []
-tool_call = true
+reasoning = false
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2023-12"
 open_weights = true
 
 [cost]
-input = 0.116
-output = 0.358
+input = 0.129
+output = 0.399
 
 [limit]
 context = 128_000
diff --git a/providers/cortecs/models/holo2-30b-a3b.toml b/providers/cortecs/models/holo2-30b-a3b.toml
new file mode 100644
index 0000000000..ac14a7e23d
--- /dev/null
+++ b/providers/cortecs/models/holo2-30b-a3b.toml
@@ -0,0 +1,23 @@
+name = "holo2-30b-a3b"
+description = "Holo2 30B A3B is a text-and-vision model optimized for analyzing graphical user interfaces, including web, desktop, and mobile, and enabling agents to interpret interfaces, reason over content, and take actions."
+release_date = "2025-12-10"
+last_updated = "2025-12-10"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.334
+output = 0.78
+
+[limit]
+context = 22_000
+output = 22_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/hy3.toml b/providers/cortecs/models/hy3.toml
deleted file mode 100644
index 994f8ca2dc..0000000000
--- a/providers/cortecs/models/hy3.toml
+++ /dev/null
@@ -1,17 +0,0 @@
-base_model = "tencent/hy3"
-# Cortecs Chat maps OpenAI `reasoning_effort`; Hy3 accepts none/low/high
-# (HF: no_think/low/high via chat_template_kwargs.reasoning_effort).
-# Cortecs API prices in EUR (0.359/0.898/0.090); converted to USD/1M at
-# EURUSD 1.1418 (2026-07-21).
-# https://api.cortecs.ai/v1/models (accessed 2026-07-22)
-# https://huggingface.co/tencent/Hy3
-structured_output = true
-reasoning_options = [{ type = "effort", values = ["none", "low", "high"] }]
-
-[cost]
-input = 0.410
-output = 1.025
-cache_read = 0.103
-
-[limit]
-context = 262_144
diff --git a/providers/cortecs/models/intellect-3.toml b/providers/cortecs/models/intellect-3.toml
deleted file mode 100644
index 207aeffa2b..0000000000
--- a/providers/cortecs/models/intellect-3.toml
+++ /dev/null
@@ -1,24 +0,0 @@
-name = "INTELLECT 3"
-description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use"
-release_date = "2025-11-26"
-last_updated = "2025-11-26"
-knowledge = "2025-11"
-attachment = true
-reasoning = true
-reasoning_options = []
-tool_call = true
-temperature = true
-open_weights = true
-
-
-[cost]
-input = 0.219
-output = 1.202
-
-[limit]
-context = 128_000
-output = 128_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/kimi-k2-instruct.toml b/providers/cortecs/models/kimi-k2-instruct.toml
deleted file mode 100644
index 2547be7ea5..0000000000
--- a/providers/cortecs/models/kimi-k2-instruct.toml
+++ /dev/null
@@ -1,24 +0,0 @@
-name = "Kimi K2 Instruct"
-description = "Kimi model for long-context chat, coding, and agentic reasoning"
-family = "kimi-k2"
-release_date = "2025-07-11"
-last_updated = "2025-09-05"
-knowledge = "2024-07"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
-
-[cost]
-input = 0.551
-output = 2.646
-
-[limit]
-context = 131_000
-output = 131_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
\ No newline at end of file
diff --git a/providers/cortecs/models/kimi-k2-thinking.toml b/providers/cortecs/models/kimi-k2-thinking.toml
deleted file mode 100644
index c3956e9045..0000000000
--- a/providers/cortecs/models/kimi-k2-thinking.toml
+++ /dev/null
@@ -1,26 +0,0 @@
-name = "Kimi K2 Thinking"
-description = "Kimi reasoning model for long-horizon research, planning, and tool use"
-release_date = "2025-12-08"
-last_updated = "2025-12-08"
-knowledge = "2025-12"
-attachment = true
-reasoning = true
-reasoning_options = []
-tool_call = true
-temperature = true
-open_weights = true
-
-[interleaved]
-field = "reasoning_content"
-
-[cost]
-input = 0.656
-output = 2.731
-
-[limit]
-context = 262_000
-output = 262_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/kimi-k2.5.toml b/providers/cortecs/models/kimi-k2.5.toml
index 29a135fa08..6e9672dcb6 100644
--- a/providers/cortecs/models/kimi-k2.5.toml
+++ b/providers/cortecs/models/kimi-k2.5.toml
@@ -1,29 +1,20 @@
-name = "Kimi K2.5"
+base_model = "moonshotai/kimi-k2.5"
 description = "Kimi reasoning model for long-horizon research, planning, and tool use"
-# Cortecs maps `thinking.type = enabled|disabled`; no effort or budget is listed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "kimi-thinking"
-release_date = "2026-01-27"
-last_updated = "2026-01-27"
-knowledge = "2025-01"
-attachment = false
-reasoning = true
-reasoning_options = [{ type = "toggle" }]
 temperature = true
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[[reasoning_options]]
+type = "toggle"
 
 [cost]
-input = 0.55
-output = 2.76
+input = 0.495
+output = 2.768
+cache_read = 0.124
 
 [limit]
-context = 256_000
 output = 256_000
 
 [modalities]
-input = ["text", "image", "video"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
+input = ["text", "image"]
diff --git a/providers/cortecs/models/kimi-k2.6.toml b/providers/cortecs/models/kimi-k2.6.toml
index b8f13eee27..629e82463b 100644
--- a/providers/cortecs/models/kimi-k2.6.toml
+++ b/providers/cortecs/models/kimi-k2.6.toml
@@ -1,29 +1,19 @@
-name = "Kimi K2.6"
+base_model = "moonshotai/kimi-k2.6"
 description = "Kimi reasoning model for long-horizon research, planning, and tool use"
-# Cortecs maps `thinking.type = enabled|disabled`; no effort or budget is listed.
-# https://api.cortecs.ai/v1/models (accessed 2026-06-25)
-family = "kimi-thinking"
-release_date = "2026-04-17"
-last_updated = "2026-04-17"
-attachment = false
-reasoning = true
-reasoning_options = [{ type = "toggle" }]
-temperature = true
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
+
+[[reasoning_options]]
+type = "toggle"
 
 [cost]
-cache_read = 0.20
-input = 0.81
-output = 3.54
+input = 0.773
+output = 3.38
+cache_read = 0.193
 
 [limit]
-context = 256_000
 output = 256_000
 
 [modalities]
 input = ["text", "image"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/kimi-k2.7-code.toml b/providers/cortecs/models/kimi-k2.7-code.toml
index 653caab5e8..77671f2c3d 100644
--- a/providers/cortecs/models/kimi-k2.7-code.toml
+++ b/providers/cortecs/models/kimi-k2.7-code.toml
@@ -5,6 +5,9 @@ reasoning_options = []
 field = "reasoning_content"
 
 [cost]
-input = 1.28
-output = 4.63
-cache_read = 0.32
+input = 0.75
+output = 3.5
+cache_read = 0.201
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/kimi-k3.toml b/providers/cortecs/models/kimi-k3.toml
index b362c683a8..002c2338ef 100644
--- a/providers/cortecs/models/kimi-k3.toml
+++ b/providers/cortecs/models/kimi-k3.toml
@@ -9,7 +9,10 @@ field = "reasoning_content"
 
 [cost]
 input = 3
-output = 15
+output = 14.999
+
+[limit]
+output = 1_048_576
 
 [modalities]
 input = ["text", "image"]
diff --git a/providers/cortecs/models/llama-3.1-405b-instruct.toml b/providers/cortecs/models/llama-3.1-405b-instruct.toml
index c25f6cc458..cc85efa5de 100644
--- a/providers/cortecs/models/llama-3.1-405b-instruct.toml
+++ b/providers/cortecs/models/llama-3.1-405b-instruct.toml
@@ -3,17 +3,18 @@ description = "Open Llama instruction model for multilingual chat, reasoning, an
 family = "llama"
 release_date = "2024-07-23"
 last_updated = "2024-07-23"
-knowledge = "2023-12"
 attachment = false
-reasoning = false
-tool_call = true
+reasoning = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2023-12"
 open_weights = true
-
+reasoning_options = []
 
 [cost]
-input = 0
-output = 0
+input = 1.95
+output = 1.95
 
 [limit]
 context = 128_000
@@ -21,4 +22,4 @@ output = 128_000
 
 [modalities]
 input = ["text"]
-output = ["text"]
\ No newline at end of file
+output = ["text"]
diff --git a/providers/cortecs/models/llama-3.1-8b-instruct.toml b/providers/cortecs/models/llama-3.1-8b-instruct.toml
new file mode 100644
index 0000000000..7b6ad92c78
--- /dev/null
+++ b/providers/cortecs/models/llama-3.1-8b-instruct.toml
@@ -0,0 +1,23 @@
+name = "llama-3.1-8b-instruct"
+description = "Optimized for dialogue, this LLM by Meta outperforms other open-source chat models in benchmarks while prioritizing helpfulness and safety."
+release_date = "2024-04-09"
+last_updated = "2024-04-09"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.167
+output = 0.167
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/llama-3.1-nemotron-ultra-253b-v1.toml b/providers/cortecs/models/llama-3.1-nemotron-ultra-253b-v1.toml
new file mode 100644
index 0000000000..7b81f53f5b
--- /dev/null
+++ b/providers/cortecs/models/llama-3.1-nemotron-ultra-253b-v1.toml
@@ -0,0 +1,23 @@
+name = "llama-3.1-nemotron-ultra-253b-v1"
+description = "A reasoning-optimized LLM based on Llama 3.1, Nemotron Ultra 253B delivers strong performance in tasks like RAG and tool use, with high efficiency and reduced latency."
+release_date = "2025-04-07"
+last_updated = "2025-04-07"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.598
+output = 1.794
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/llama-3.3-70b-instruct.toml b/providers/cortecs/models/llama-3.3-70b-instruct.toml
index 0074929273..cb054b806d 100644
--- a/providers/cortecs/models/llama-3.3-70b-instruct.toml
+++ b/providers/cortecs/models/llama-3.3-70b-instruct.toml
@@ -1,12 +1,10 @@
 base_model = "meta/llama-3.3-70b-instruct"
-name = "Llama 3.3 70B Instruct"
 attachment = false
-reasoning = true
-reasoning_options = []
+structured_output = true
 
 [cost]
-input = 0.089
-output = 0.275
+input = 0.129
+output = 0.399
 
 [limit]
 context = 131_000
diff --git a/providers/cortecs/models/llama-4-maverick.toml b/providers/cortecs/models/llama-4-maverick.toml
deleted file mode 100644
index cec8b806b3..0000000000
--- a/providers/cortecs/models/llama-4-maverick.toml
+++ /dev/null
@@ -1,7 +0,0 @@
-base_model = "meta/llama-4-maverick-17b-instruct"
-
-[cost]
-input = 0.124
-output = 0.603
-cache_read = 0.03
-cache_write = 0.151
\ No newline at end of file
diff --git a/providers/cortecs/models/minicpm-v-4.5.toml b/providers/cortecs/models/minicpm-v-4.5.toml
new file mode 100644
index 0000000000..1ec22b6a6e
--- /dev/null
+++ b/providers/cortecs/models/minicpm-v-4.5.toml
@@ -0,0 +1,22 @@
+name = "minicpm-v-4.5"
+description = "MiniCPM-V 4.5 is a compact, high-performance vision-language model excelling in video understanding, OCR, and multimodal reasoning with efficient deployment."
+release_date = "2026-06-02"
+last_updated = "2026-06-02"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.651
+output = 1.097
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/minimax-m2.1.toml b/providers/cortecs/models/minimax-m2.1.toml
index d75abf7f9f..1e3b16ea6e 100644
--- a/providers/cortecs/models/minimax-m2.1.toml
+++ b/providers/cortecs/models/minimax-m2.1.toml
@@ -1,26 +1,15 @@
-name = "MiniMax-M2.1"
+base_model = "minimax/MiniMax-M2.1"
 description = "MiniMax model for chat, coding, office work, and agentic tasks"
-family = "minimax"
-release_date = "2025-12-23"
-last_updated = "2025-12-23"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
 
 [cost]
-input = 0.34
-output = 1.34
+input = 0.359
+output = 1.435
 
 [limit]
 context = 196_000
 output = 196_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/minimax-m2.5.toml b/providers/cortecs/models/minimax-m2.5.toml
index 6518f06631..1f41647ea4 100644
--- a/providers/cortecs/models/minimax-m2.5.toml
+++ b/providers/cortecs/models/minimax-m2.5.toml
@@ -1,26 +1,16 @@
-name = "MiniMax-M2.5"
+base_model = "minimax/MiniMax-M2.5"
 description = "MiniMax model for chat, coding, office work, and agentic tasks"
-family = "minimax"
-release_date = "2026-02-12"
-last_updated = "2026-02-12"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
 
 [cost]
-input = 0.32
-output = 1.18
+input = 0.296
+output = 1.087
+cache_read = 0.03
 
 [limit]
-context = 196_608
+context = 196_680
 output = 196_608
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/minimax-m2.7.toml b/providers/cortecs/models/minimax-m2.7.toml
index 25badaaa22..de16979353 100644
--- a/providers/cortecs/models/minimax-m2.7.toml
+++ b/providers/cortecs/models/minimax-m2.7.toml
@@ -1,24 +1,12 @@
-name = "MiniMax-m2.7"
+base_model = "minimax/MiniMax-M2.7"
 description = "MiniMax model for chat, coding, office work, and agentic tasks"
-family = "minimax"
-release_date = "2026-03-18"
-last_updated = "2026-03-18"
-attachment = false
-reasoning = true
-reasoning_options = []
-temperature = true
-tool_call = true
 structured_output = true
-open_weights = true
+reasoning_options = []
 
 [cost]
-input = 0.47
-output = 1.40
+input = 0.668
+output = 2.674
 
 [limit]
-context = 202_752
+context = 196_608
 output = 196_072
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/minimax-m2.toml b/providers/cortecs/models/minimax-m2.toml
index 11cf0ae566..6e428e287e 100644
--- a/providers/cortecs/models/minimax-m2.toml
+++ b/providers/cortecs/models/minimax-m2.toml
@@ -1,27 +1,15 @@
-name = "MiniMax-M2"
+base_model = "minimax/MiniMax-M2"
 description = "MiniMax model for chat, coding, office work, and agentic tasks"
-family = "minimax"
-release_date = "2025-10-27"
-last_updated = "2025-10-27"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-temperature = true
-knowledge = "2024-11"
-tool_call = true
-open_weights = true
+
+[interleaved]
+field = "reasoning_content"
 
 [cost]
-input = 0.39
-output = 1.57
+input = 0.349
+output = 1.405
 
 [limit]
 context = 400_000
 output = 400_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
-
-[interleaved]
-field = "reasoning_content"
diff --git a/providers/cortecs/models/minimax-m3.toml b/providers/cortecs/models/minimax-m3.toml
index b5f5f34f91..8201c1a880 100644
--- a/providers/cortecs/models/minimax-m3.toml
+++ b/providers/cortecs/models/minimax-m3.toml
@@ -1,7 +1,15 @@
 base_model = "minimax/MiniMax-M3"
+structured_output = true
 reasoning_options = []
 
 [cost]
-input = 0.355
-output = 1.775
-cache_read = 0.089
+input = 0.395
+output = 1.977
+cache_read = 0.099
+
+[limit]
+context = 1_048_576
+output = 1_048_576
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/ministral-14b-2512.toml b/providers/cortecs/models/ministral-14b-2512.toml
new file mode 100644
index 0000000000..e57017e1d7
--- /dev/null
+++ b/providers/cortecs/models/ministral-14b-2512.toml
@@ -0,0 +1,23 @@
+name = "ministral-14b-2512"
+description = "Ministral 3 14B is a frontier-level 14B multimodal model optimized for local deployment, delivering state-of-the-art text and vision reasoning with a 256K context window and strong agentic capabilities."
+release_date = "2025-12-03"
+last_updated = "2025-12-03"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.223
+output = 0.223
+cache_read = 0.022
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/ministral-3b-2512.toml b/providers/cortecs/models/ministral-3b-2512.toml
new file mode 100644
index 0000000000..4e686e35de
--- /dev/null
+++ b/providers/cortecs/models/ministral-3b-2512.toml
@@ -0,0 +1,23 @@
+name = "ministral-3b-2512"
+description = "Ministral 3 3B is a compact, efficient multimodal model with strong language, vision capabilities, and ideal for custom fine-tuning."
+release_date = "2025-12-03"
+last_updated = "2025-12-03"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.111
+output = 0.111
+cache_read = 0.011
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/ministral-8b-2512.toml b/providers/cortecs/models/ministral-8b-2512.toml
new file mode 100644
index 0000000000..ddbd962448
--- /dev/null
+++ b/providers/cortecs/models/ministral-8b-2512.toml
@@ -0,0 +1,23 @@
+name = "ministral-8b-2512"
+description = "Ministral 3 8B is a balanced, efficient multimodal model offering strong text and vision capabilities, optimized for edge and local deployment."
+release_date = "2025-12-03"
+last_updated = "2025-12-03"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.167
+output = 0.167
+cache_read = 0.017
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-7b-instruct-v0.2.toml b/providers/cortecs/models/mistral-7b-instruct-v0.2.toml
new file mode 100644
index 0000000000..eeb7abd05e
--- /dev/null
+++ b/providers/cortecs/models/mistral-7b-instruct-v0.2.toml
@@ -0,0 +1,22 @@
+name = "mistral-7b-instruct-v0.2"
+description = "Mistral 7B Instruct is a compact, 7B parameter model optimized for fast and efficient text and code generation with a 32K token context window."
+release_date = "2025-05-26"
+last_updated = "2025-05-26"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0.159
+output = 0.219
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-7b-instruct-v0.3.toml b/providers/cortecs/models/mistral-7b-instruct-v0.3.toml
new file mode 100644
index 0000000000..9e67e8f964
--- /dev/null
+++ b/providers/cortecs/models/mistral-7b-instruct-v0.3.toml
@@ -0,0 +1,22 @@
+name = "mistral-7b-instruct-v0.3"
+description = "Mistral-7B-Instruct-v0.3 model is a fine-tuned version of the Mistral 7B base model, optimized for instruction-following tasks. Released in 2023, it is intended for demonstration purposes and does not include built-in guardrails or moderation features."
+release_date = "2025-05-26"
+last_updated = "2025-05-26"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.111
+output = 0.111
+
+[limit]
+context = 127_000
+output = 127_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-large-2402.toml b/providers/cortecs/models/mistral-large-2402.toml
new file mode 100644
index 0000000000..5e13b1f6ca
--- /dev/null
+++ b/providers/cortecs/models/mistral-large-2402.toml
@@ -0,0 +1,23 @@
+name = "mistral-large-2402"
+description = "Mistral Large (24.02) is Mistral AI’s most advanced language model, built for complex multilingual reasoning, code generation, and deep text understanding."
+release_date = "2025-05-26"
+last_updated = "2025-05-26"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 4.284
+output = 12.952
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-large-2512.toml b/providers/cortecs/models/mistral-large-2512.toml
index ec9326edc8..52b4e581a2 100644
--- a/providers/cortecs/models/mistral-large-2512.toml
+++ b/providers/cortecs/models/mistral-large-2512.toml
@@ -1,13 +1,10 @@
 base_model = "mistral/mistral-large-2512"
-name = "Mistral Large 3 2512"
-release_date = "2025-12-01"
-last_updated = "2025-12-01"
-knowledge = "2025-12"
+structured_output = true
 
 [cost]
-input = 0.5
-output = 1.5
-cache_read = 0.05
+input = 0.557
+output = 1.671
+cache_read = 0.056
 
 [limit]
 context = 256_000
diff --git a/providers/cortecs/models/mistral-medium-2508.toml b/providers/cortecs/models/mistral-medium-2508.toml
new file mode 100644
index 0000000000..e81a6c42e6
--- /dev/null
+++ b/providers/cortecs/models/mistral-medium-2508.toml
@@ -0,0 +1,24 @@
+name = "mistral-medium-2508"
+description = "Mistral Medium 2508 is a frontier-class multimodal LLM with a 128,000 token context window, optimized for reasoning, coding, and multimodal tasks."
+release_date = "2024-08-07"
+last_updated = "2024-08-07"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.446
+output = 2.228
+cache_read = 0.045
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-medium-3.5.toml b/providers/cortecs/models/mistral-medium-3.5.toml
new file mode 100644
index 0000000000..c930be77d5
--- /dev/null
+++ b/providers/cortecs/models/mistral-medium-3.5.toml
@@ -0,0 +1,23 @@
+name = "mistral-medium-3.5"
+description = "Mistral Medium 3.5 is a frontier multimodal 128B model combining reasoning, coding, and instruction-following with strong agentic performance and efficient deployment."
+release_date = "2026-04-30"
+last_updated = "2026-04-30"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 1.671
+output = 5.57
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-nemo-instruct-2407.toml b/providers/cortecs/models/mistral-nemo-instruct-2407.toml
new file mode 100644
index 0000000000..6b38bfabf3
--- /dev/null
+++ b/providers/cortecs/models/mistral-nemo-instruct-2407.toml
@@ -0,0 +1,23 @@
+name = "mistral-nemo-instruct-2407"
+description = "A 12B parameter, instruct-tuned language model by Mistral AI and NVIDIA, designed for advanced instruction following, multi-turn conversations, and generating text and code across multiple languages."
+release_date = "2024-08-07"
+last_updated = "2024-08-07"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.145
+output = 0.145
+cache_read = 0.014
+
+[limit]
+context = 128_000
+output = 131_072
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-small-2503.toml b/providers/cortecs/models/mistral-small-2503.toml
new file mode 100644
index 0000000000..3a21b6e50e
--- /dev/null
+++ b/providers/cortecs/models/mistral-small-2503.toml
@@ -0,0 +1,22 @@
+name = "mistral-small-2503"
+description = "Combines advanced text and vision capabilities with 24 billion parameters, supporting multilingual tasks and long contexts up to 131k tokens, making it versatile for various applications without sacrificing performance."
+release_date = "2025-03-20"
+last_updated = "2025-03-20"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.111
+output = 0.334
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/mistral-small-2603.toml b/providers/cortecs/models/mistral-small-2603.toml
new file mode 100644
index 0000000000..4ec4973486
--- /dev/null
+++ b/providers/cortecs/models/mistral-small-2603.toml
@@ -0,0 +1,12 @@
+base_model = "mistral/mistral-small-2603"
+structured_output = true
+reasoning_options = []
+
+[cost]
+input = 0.143
+output = 0.568
+cache_read = 0.014
+
+[limit]
+context = 262_144
+output = 262_144
diff --git a/providers/cortecs/models/mistral-small-3.2-24b-instruct-2506.toml b/providers/cortecs/models/mistral-small-3.2-24b-instruct-2506.toml
new file mode 100644
index 0000000000..e98e612030
--- /dev/null
+++ b/providers/cortecs/models/mistral-small-3.2-24b-instruct-2506.toml
@@ -0,0 +1,22 @@
+name = "mistral-small-3.2-24b-instruct-2506"
+description = "Mistral-Small-3.2-24B-Instruct-2506 is a 24B parameter instruction-tuned model with enhanced long-context support (128k) and state-of-the-art vision understanding."
+release_date = "2025-05-26"
+last_updated = "2025-05-26"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.1
+output = 0.312
+
+[limit]
+context = 131_000
+output = 131_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/mixtral-8x7B-instruct-v0.1.toml b/providers/cortecs/models/mixtral-8x7B-instruct-v0.1.toml
index 21aff81407..f967e6241e 100644
--- a/providers/cortecs/models/mixtral-8x7B-instruct-v0.1.toml
+++ b/providers/cortecs/models/mixtral-8x7B-instruct-v0.1.toml
@@ -2,17 +2,18 @@ name = "Mixtral 8x7B Instruct v0.1"
 description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use"
 release_date = "2023-12-11"
 last_updated = "2023-12-11"
-knowledge = "2023-09"
 attachment = false
 reasoning = true
-reasoning_options = []
-tool_call = false
 temperature = true
+tool_call = false
+structured_output = true
+knowledge = "2023-09"
 open_weights = true
+reasoning_options = []
 
 [cost]
-input = 0.438
-output = 0.68
+input = 0.488
+output = 0.758
 
 [limit]
 context = 32_000
diff --git a/providers/cortecs/models/nemotron-3-super-120b-a12b.toml b/providers/cortecs/models/nemotron-3-super-120b-a12b.toml
deleted file mode 100644
index 43f86fcf83..0000000000
--- a/providers/cortecs/models/nemotron-3-super-120b-a12b.toml
+++ /dev/null
@@ -1,23 +0,0 @@
-name = "Nemotron 3 Super 120B A12B"
-base_model = "nvidia/nemotron-3-super-120b-a12b"
-release_date = "2026-03-11"
-last_updated = "2026-03-11"
-knowledge = "2025-12"
-attachment = false
-reasoning = true
-reasoning_options = []
-tool_call = true
-temperature = true
-open_weights = true
-
-[cost]
-input = 0.266
-output = 0.799
-
-[limit]
-context = 262_144
-output = 262_144
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/nemotron-nano-v2-12b.toml b/providers/cortecs/models/nemotron-nano-v2-12b.toml
new file mode 100644
index 0000000000..017dad0bf3
--- /dev/null
+++ b/providers/cortecs/models/nemotron-nano-v2-12b.toml
@@ -0,0 +1,23 @@
+name = "nemotron-nano-v2-12b"
+description = "NVIDIA Nemotron Nano v2 12B is a 12-billion-parameter multimodal reasoning model designed for advanced video understanding, document intelligence, and visual reasoning, built with a hybrid Transformer-Mamba architecture for high efficiency and low latency."
+release_date = "2025-10-31"
+last_updated = "2025-10-31"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.24
+output = 0.707
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/nova-2-lite.toml b/providers/cortecs/models/nova-2-lite.toml
new file mode 100644
index 0000000000..b0b23bfad4
--- /dev/null
+++ b/providers/cortecs/models/nova-2-lite.toml
@@ -0,0 +1,23 @@
+name = "nova-2-lite"
+description = "Nova 2 Lite is an advanced multimodal reasoning model that combines efficiency and performance, delivering reliable AI for agentic workflows and enterprise applications."
+release_date = "2025-12-04"
+last_updated = "2025-12-04"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.373
+output = 3.144
+
+[limit]
+context = 1_000_000
+output = 1_000_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/nova-lite-v1.toml b/providers/cortecs/models/nova-lite-v1.toml
new file mode 100644
index 0000000000..a0199cc23d
--- /dev/null
+++ b/providers/cortecs/models/nova-lite-v1.toml
@@ -0,0 +1,23 @@
+name = "nova-lite-v1"
+description = "Nova Lite is a fast, low-cost multimodal foundation model capable of reasoning over text, images, and video in 200+ languages."
+release_date = "2025-04-14"
+last_updated = "2025-04-14"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.069
+output = 0.275
+
+[limit]
+context = 300_000
+output = 300_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/nova-micro-v1.toml b/providers/cortecs/models/nova-micro-v1.toml
new file mode 100644
index 0000000000..ce8595fb24
--- /dev/null
+++ b/providers/cortecs/models/nova-micro-v1.toml
@@ -0,0 +1,23 @@
+name = "nova-micro-v1"
+description = "Nova Micro is a multilingual text-to-text foundation model with strong reasoning capabilities and broad language coverage across 200+ languages."
+release_date = "2025-04-14"
+last_updated = "2025-04-14"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.04
+output = 0.159
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/nova-pro-v1.toml b/providers/cortecs/models/nova-pro-v1.toml
index c66e5e3add..09524c68f3 100644
--- a/providers/cortecs/models/nova-pro-v1.toml
+++ b/providers/cortecs/models/nova-pro-v1.toml
@@ -3,17 +3,18 @@ description = "Flagship model for demanding analysis, coding, and production age
 family = "nova-pro"
 release_date = "2024-12-03"
 last_updated = "2024-12-03"
-knowledge = "2024-04"
-attachment = false
-reasoning = false
-tool_call = true
+attachment = true
+reasoning = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2024-04"
 open_weights = false
-
+reasoning_options = []
 
 [cost]
-input = 1.016
-output = 4.061
+input = 0.918
+output = 3.671
 
 [limit]
 context = 300_000
@@ -21,4 +22,4 @@ output = 5_000
 
 [modalities]
 input = ["text", "image"]
-output = ["text"]
\ No newline at end of file
+output = ["text"]
diff --git a/providers/cortecs/models/nvidia-nemotron-3-nano-30b-a3b.toml b/providers/cortecs/models/nvidia-nemotron-3-nano-30b-a3b.toml
new file mode 100644
index 0000000000..d1ed1aa83e
--- /dev/null
+++ b/providers/cortecs/models/nvidia-nemotron-3-nano-30b-a3b.toml
@@ -0,0 +1,23 @@
+name = "nvidia-nemotron-3-nano-30b-a3b"
+description = "Nemotron-Nano-3-30B-A3B is a compact Mixture-of-Experts model optimized for efficient reasoning, chat, and coding, with strong multilingual support and long-context RAG and agent workflows."
+release_date = "2026-01-12"
+last_updated = "2026-01-12"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.06
+output = 0.24
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/nvidia-nemotron-3-nano-omni.toml b/providers/cortecs/models/nvidia-nemotron-3-nano-omni.toml
new file mode 100644
index 0000000000..4b8a70d55a
--- /dev/null
+++ b/providers/cortecs/models/nvidia-nemotron-3-nano-omni.toml
@@ -0,0 +1,23 @@
+name = "nvidia-nemotron-3-nano-omni"
+description = "Nemotron-3-Nano-Omni is an open, efficient omni-modal reasoning model that unifies text, image, audio, and video for agentic AI workflows."
+release_date = "2026-04-29"
+last_updated = "2026-04-29"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.059
+output = 0.237
+
+[limit]
+context = 300_000
+output = 300_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/pixtral-12b-2409.toml b/providers/cortecs/models/pixtral-12b-2409.toml
new file mode 100644
index 0000000000..62a978e522
--- /dev/null
+++ b/providers/cortecs/models/pixtral-12b-2409.toml
@@ -0,0 +1,23 @@
+name = "pixtral-12b-2409"
+description = "Pixtral 2409 12B is a state-of-the-art multimodal model with 12B parameters and a 400M vision encoder, natively trained on interleaved text and image data. It excels in tasks spanning vision-language reasoning, instruction following, and pure text understanding, making it highly effective for real-world multimodal applications."
+release_date = "2024-11-09"
+last_updated = "2024-11-09"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.223
+output = 0.223
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/pixtral-large-2502.toml b/providers/cortecs/models/pixtral-large-2502.toml
new file mode 100644
index 0000000000..2e2fb46834
--- /dev/null
+++ b/providers/cortecs/models/pixtral-large-2502.toml
@@ -0,0 +1,23 @@
+name = "pixtral-large-2502"
+description = "Pixtral Large (25.02) is a 124B open-weight multimodal model built on Mistral Large 2, offering advanced image understanding and strong performance across text and code tasks."
+release_date = "2025-05-26"
+last_updated = "2025-05-26"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 1.993
+output = 5.978
+
+[limit]
+context = 128_000
+output = 128_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/qwen-2.5-72b-instruct.toml b/providers/cortecs/models/qwen-2.5-72b-instruct.toml
deleted file mode 100644
index d22e8a2ba0..0000000000
--- a/providers/cortecs/models/qwen-2.5-72b-instruct.toml
+++ /dev/null
@@ -1,23 +0,0 @@
-name = "Qwen2.5 72B Instruct"
-description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
-family = "qwen"
-release_date = "2024-09-19"
-last_updated = "2024-09-19"
-knowledge = "2024-06"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
-[cost]
-input = 0.062
-output = 0.231
-
-[limit]
-context = 33_000
-output = 33_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/qwen2.5-vl-72b-instruct.toml b/providers/cortecs/models/qwen2.5-vl-72b-instruct.toml
new file mode 100644
index 0000000000..463981d633
--- /dev/null
+++ b/providers/cortecs/models/qwen2.5-vl-72b-instruct.toml
@@ -0,0 +1,23 @@
+name = "qwen2.5-vl-72b-instruct"
+description = "Qwen2.5-VL is a powerful vision-language model with advanced capabilities in visual understanding, long video reasoning, and structured output generation."
+release_date = "2025-01-27"
+last_updated = "2025-01-27"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = false
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.25
+output = 0.747
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/qwen3-235b-a22b-instruct-2507.toml b/providers/cortecs/models/qwen3-235b-a22b-instruct-2507.toml
index bd49acb488..cc19a30734 100644
--- a/providers/cortecs/models/qwen3-235b-a22b-instruct-2507.toml
+++ b/providers/cortecs/models/qwen3-235b-a22b-instruct-2507.toml
@@ -3,20 +3,22 @@ description = "Qwen instruction model for multilingual chat, reasoning, and tool
 family = "qwen"
 release_date = "2025-07-23"
 last_updated = "2025-07-23"
-knowledge = "2025-04"
 attachment = false
 reasoning = true
-reasoning_options = []
-tool_call = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2025-04"
 open_weights = true
+reasoning_options = []
 
 [cost]
-input = 0.062
-output = 0.408
+input = 0.069
+output = 0.455
+cache_read = 0.018
 
 [limit]
-context = 131_000
+context = 262_000
 output = 131_000
 
 [modalities]
diff --git a/providers/cortecs/models/qwen3-30b-a3b-instruct-2507.toml b/providers/cortecs/models/qwen3-30b-a3b-instruct-2507.toml
new file mode 100644
index 0000000000..3b44b48f5b
--- /dev/null
+++ b/providers/cortecs/models/qwen3-30b-a3b-instruct-2507.toml
@@ -0,0 +1,23 @@
+name = "qwen3-30b-a3b-instruct-2507"
+description = "Qwen3-30B-A3B-Instruct-2507 is an advanced Mixture-of-Experts model optimized for reasoning, coding, and multilingual instruction following."
+release_date = "2025-07-28"
+last_updated = "2025-07-28"
+attachment = false
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.099
+output = 0.299
+
+[limit]
+context = 262_000
+output = 262_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/qwen3-32b.toml b/providers/cortecs/models/qwen3-32b.toml
index 4d795ca6b2..23c4b00ad5 100644
--- a/providers/cortecs/models/qwen3-32b.toml
+++ b/providers/cortecs/models/qwen3-32b.toml
@@ -1,24 +1,12 @@
-name = "Qwen3 32B"
+base_model = "alibaba/qwen3-32b"
 description = "Qwen instruction model for multilingual chat, reasoning, and tool use"
-family = "qwen"
-release_date = "2025-04-29"
-last_updated = "2025-04-29"
-knowledge = "2024-12"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
+structured_output = true
+reasoning_options = []
 
 [cost]
 input = 0.099
-output = 0.33
+output = 0.299
 
 [limit]
-context = 16_384
-output = 16_384
-
-[modalities]
-input = ["text"]
-output = ["text"]
\ No newline at end of file
+context = 40_000
+output = 40_000
diff --git a/providers/cortecs/models/qwen3-coder-30b-a3b-instruct.toml b/providers/cortecs/models/qwen3-coder-30b-a3b-instruct.toml
index 9ce13fc845..09bb3833a2 100644
--- a/providers/cortecs/models/qwen3-coder-30b-a3b-instruct.toml
+++ b/providers/cortecs/models/qwen3-coder-30b-a3b-instruct.toml
@@ -1,14 +1,10 @@
 base_model = "alibaba/qwen3-coder-30b-a3b-instruct"
-name = "Qwen3 Coder 30B A3B Instruct"
-release_date = "2025-07-31"
-last_updated = "2025-07-31"
-reasoning = true
-reasoning_options = []
+structured_output = true
 
 [cost]
-input = 0.053
-output = 0.222
+input = 0.067
+output = 0.245
+cache_read = 0.014
 
 [limit]
-context = 262_000
 output = 262_000
diff --git a/providers/cortecs/models/qwen3-coder-480b-a35b-instruct.toml b/providers/cortecs/models/qwen3-coder-480b-a35b-instruct.toml
deleted file mode 100644
index fde370bc12..0000000000
--- a/providers/cortecs/models/qwen3-coder-480b-a35b-instruct.toml
+++ /dev/null
@@ -1,24 +0,0 @@
-name = "Qwen3 Coder 480B A35B Instruct"
-description = "Qwen coding model for software agents, repository edits, and code reasoning"
-family = "qwen"
-release_date = "2025-07-25"
-last_updated = "2025-07-25"
-knowledge = "2025-01"
-attachment = false
-reasoning = false
-tool_call = true
-temperature = true
-open_weights = true
-
-
-[cost]
-input = 0.441
-output = 1.984
-
-[limit]
-context = 262_000
-output = 262_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
\ No newline at end of file
diff --git a/providers/cortecs/models/qwen3-coder-next.toml b/providers/cortecs/models/qwen3-coder-next.toml
index 2b3c74cb78..36f18ead84 100644
--- a/providers/cortecs/models/qwen3-coder-next.toml
+++ b/providers/cortecs/models/qwen3-coder-next.toml
@@ -3,17 +3,18 @@ description = "Qwen coding model for software agents, repository edits, and code
 family = "qwen"
 release_date = "2026-02-04"
 last_updated = "2026-02-04"
-knowledge = "2025-04"
 attachment = false
 reasoning = true
-reasoning_options = []
-tool_call = true
 temperature = true
+tool_call = true
+structured_output = true
+knowledge = "2025-04"
 open_weights = true
+reasoning_options = []
 
 [cost]
-input = 0.158
-output = 0.84
+input = 0.167
+output = 0.891
 
 [limit]
 context = 256_000
diff --git a/providers/cortecs/models/qwen3-next-80b-a3b-thinking.toml b/providers/cortecs/models/qwen3-next-80b-a3b-thinking.toml
index dd1b63f45d..d0add520f9 100644
--- a/providers/cortecs/models/qwen3-next-80b-a3b-thinking.toml
+++ b/providers/cortecs/models/qwen3-next-80b-a3b-thinking.toml
@@ -1,24 +1,12 @@
-name = "Qwen3 Next 80B A3B Thinking"
+base_model = "alibaba/qwen3-next-80b-a3b-thinking"
 description = "Reasoning model for deliberate analysis, multi-step problem solving, and tool use"
-release_date = "2025-09-11"
-last_updated = "2025-09-11"
-knowledge = "2025-04"
-attachment = false
-reasoning = true
+structured_output = true
 reasoning_options = []
-tool_call = true
-temperature = true
-open_weights = true
-
 
 [cost]
-input = 0.164
-output = 1.311
+input = 0.149
+output = 1.195
 
 [limit]
 context = 128_000
 output = 128_000
-
-[modalities]
-input = ["text"]
-output = ["text"]
diff --git a/providers/cortecs/models/qwen3-vl-235b-a22b.toml b/providers/cortecs/models/qwen3-vl-235b-a22b.toml
new file mode 100644
index 0000000000..b6a11040be
--- /dev/null
+++ b/providers/cortecs/models/qwen3-vl-235b-a22b.toml
@@ -0,0 +1,24 @@
+name = "qwen3-vl-235b-a22b"
+description = "Qwen3 VL 235B A22B is a 235B-parameter MoE vision-language flagship model (≈22B active) designed for frontier-level multimodal understanding across text, images, documents, and long videos."
+release_date = "2026-01-13"
+last_updated = "2026-01-13"
+attachment = true
+reasoning = true
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+reasoning_options = []
+
+[cost]
+input = 0.617
+output = 3.119
+cache_read = 0.052
+
+[limit]
+context = 256_000
+output = 256_000
+
+[modalities]
+input = ["text", "image"]
+output = ["text"]
diff --git a/providers/cortecs/models/qwen3.5-122b-a10b.toml b/providers/cortecs/models/qwen3.5-122b-a10b.toml
index c0c7090c1e..ab646dfbd1 100644
--- a/providers/cortecs/models/qwen3.5-122b-a10b.toml
+++ b/providers/cortecs/models/qwen3.5-122b-a10b.toml
@@ -1,15 +1,12 @@
 base_model = "alibaba/qwen3.5-122b-a10b"
 base_model_omit = ["structured_output"]
-reasoning_options = []
-name = "Qwen3.5 122B A10B"
-release_date = "2026-02-24"
-last_updated = "2026-02-24"
 attachment = false
-knowledge = "2026-01"
+reasoning_options = []
 
 [cost]
-input = 0.444
-output = 3.106
+input = 0.495
+output = 3.46
+cache_read = 0.124
 
 [limit]
 output = 262_144
diff --git a/providers/cortecs/models/qwen3.5-397b-a17b.toml b/providers/cortecs/models/qwen3.5-397b-a17b.toml
index 1c8f0a0c4a..a2ce4e256d 100644
--- a/providers/cortecs/models/qwen3.5-397b-a17b.toml
+++ b/providers/cortecs/models/qwen3.5-397b-a17b.toml
@@ -1,18 +1,14 @@
 base_model = "alibaba/qwen3.5-397b-a17b"
 base_model_omit = ["structured_output"]
-reasoning_options = []
-name = "Qwen3.5 397B A17B"
-release_date = "2026-02-16"
-last_updated = "2026-02-16"
 attachment = false
-knowledge = "2026-01"
+reasoning_options = []
 
 [cost]
-input = 0.6
-output = 3.6
+input = 0.668
+output = 4.01
 
 [limit]
-context = 250_000
+context = 262_000
 output = 250_000
 
 [modalities]
diff --git a/providers/cortecs/models/qwen3.5-9b.toml b/providers/cortecs/models/qwen3.5-9b.toml
new file mode 100644
index 0000000000..9cfd9d5d5a
--- /dev/null
+++ b/providers/cortecs/models/qwen3.5-9b.toml
@@ -0,0 +1,12 @@
+base_model = "alibaba/qwen3.5-9b"
+reasoning_options = []
+
+[cost]
+input = 0.111
+output = 0.167
+
+[limit]
+output = 262_144
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/qwen3.6-27b.toml b/providers/cortecs/models/qwen3.6-27b.toml
new file mode 100644
index 0000000000..dc5b852218
--- /dev/null
+++ b/providers/cortecs/models/qwen3.6-27b.toml
@@ -0,0 +1,13 @@
+base_model = "alibaba/qwen3.6-27b"
+reasoning_options = []
+
+[cost]
+input = 0.446
+output = 3.008
+
+[limit]
+context = 262_000
+output = 262_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/qwen3.6-35b-a3b.toml b/providers/cortecs/models/qwen3.6-35b-a3b.toml
new file mode 100644
index 0000000000..2c3a0bd470
--- /dev/null
+++ b/providers/cortecs/models/qwen3.6-35b-a3b.toml
@@ -0,0 +1,13 @@
+base_model = "alibaba/qwen3.6-35b-a3b"
+reasoning_options = []
+
+[cost]
+input = 0.167
+output = 0.557
+
+[limit]
+context = 262_000
+output = 262_000
+
+[modalities]
+input = ["text", "image"]
diff --git a/providers/cortecs/models/qwen3guard-gen-0.6b.toml b/providers/cortecs/models/qwen3guard-gen-0.6b.toml
new file mode 100644
index 0000000000..e74f24fc1e
--- /dev/null
+++ b/providers/cortecs/models/qwen3guard-gen-0.6b.toml
@@ -0,0 +1,22 @@
+name = "qwen3guard-gen-0.6b"
+description = "Qwen3Guard-Gen-0.6B is a lightweight multilingual safety moderation model that classifies prompts and responses into safe, controversial, or unsafe categories."
+release_date = "2026-02-04"
+last_updated = "2026-02-04"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0
+output = 0
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/qwen3guard-gen-8b.toml b/providers/cortecs/models/qwen3guard-gen-8b.toml
new file mode 100644
index 0000000000..e749c5af1d
--- /dev/null
+++ b/providers/cortecs/models/qwen3guard-gen-8b.toml
@@ -0,0 +1,22 @@
+name = "qwen3guard-gen-8b"
+description = "Qwen3Guard-Gen-8B is a large-scale multilingual safety moderation model designed for high-accuracy prompt and response classification."
+release_date = "2026-02-04"
+last_updated = "2026-02-04"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 0
+output = 0
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text"]
+output = ["text"]
diff --git a/providers/cortecs/models/voxtral-small-2507.toml b/providers/cortecs/models/voxtral-small-2507.toml
new file mode 100644
index 0000000000..f8ce5732ff
--- /dev/null
+++ b/providers/cortecs/models/voxtral-small-2507.toml
@@ -0,0 +1,23 @@
+name = "voxtral-small-2507"
+description = "Voxtral Small is a multimodal model with audio input, combining advanced speech capabilities with strong text performance for transcription, translation, and audio understanding."
+release_date = "2026-02-02"
+last_updated = "2026-02-02"
+attachment = true
+reasoning = false
+temperature = false
+tool_call = true
+structured_output = true
+open_weights = false
+
+[cost]
+input = 0.111
+output = 0.334
+cache_read = 0.011
+
+[limit]
+context = 32_000
+output = 32_000
+
+[modalities]
+input = ["text", "audio"]
+output = ["text"]