Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 33 additions & 0 deletions providers/cafe24/logo.svg
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
11 changes: 11 additions & 0 deletions providers/cafe24/models/MiniMaxAI/MiniMax-M3.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
# KRW list price W429/W1,716 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $0.30/1.20 per MTok.
# Thinking toggle verified forwarded on 2026-08-04 via `chat_template_kwargs.thinking_mode`
# (enabled/disabled/adaptive): identical prompt yields 26 / 2 / 16 completion tokens respectively.
# OpenAI-style `reasoning_effort` and Anthropic-style `thinking` are accepted but not forwarded.
base_model = "minimax/MiniMax-M3"
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.3
output = 1.2
13 changes: 13 additions & 0 deletions providers/cafe24/models/Qwen/Qwen3-32B.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
# KRW list price W115/W401 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $0.08/0.28 per MTok.
# Thinking toggle verified forwarded on 2026-08-04 via `chat_template_kwargs.enable_thinking`:
# default(thinking)=320 completion tokens vs enable_thinking=false -> 2 tokens on an identical prompt.
base_model = "alibaba/qwen3-32b"
reasoning_options = [{ type = "toggle" }]

[cost]
input = 0.08
output = 0.28

[limit]
context = 131_000
12 changes: 12 additions & 0 deletions providers/cafe24/models/anthropic/claude-fable-5.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
# KRW list price W14,300/W71,500 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $10/50 per MTok.
# Reasoning controls verified empirically on 2026-08-04 against the live gateway:
# OpenAI-style `reasoning_effort` and Anthropic-style `thinking` are accepted but not
# forwarded (no reasoning-token accounting, no output differential) -> affirmatively no
# caller-controllable reasoning surface; hence `reasoning_options = []`.
base_model = "anthropic/claude-fable-5"
reasoning_options = []

[cost]
input = 10
output = 50
15 changes: 15 additions & 0 deletions providers/cafe24/models/anthropic/claude-haiku-4-5.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
# KRW list price W1,430/W7,150 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $1/5 per MTok.
# Reasoning controls verified empirically on 2026-08-04 against the live gateway:
# OpenAI-style `reasoning_effort` and Anthropic-style `thinking` are accepted but not
# forwarded (no reasoning-token accounting, no output differential) -> affirmatively no
# caller-controllable reasoning surface; hence `reasoning_options = []`.
base_model = "anthropic/claude-haiku-4-5"
reasoning_options = []

[cost]
input = 1
output = 5

[limit]
output = 16_384
12 changes: 12 additions & 0 deletions providers/cafe24/models/anthropic/claude-opus-5.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
# KRW list price W7,150/W35,750 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $5/25 per MTok.
# Reasoning controls verified empirically on 2026-08-04 against the live gateway:
# OpenAI-style `reasoning_effort` and Anthropic-style `thinking` are accepted but not
# forwarded (no reasoning-token accounting, no output differential) -> affirmatively no
# caller-controllable reasoning surface; hence `reasoning_options = []`.
base_model = "anthropic/claude-opus-5"
reasoning_options = []

[cost]
input = 5
output = 25
12 changes: 12 additions & 0 deletions providers/cafe24/models/anthropic/claude-sonnet-5.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
# KRW list price W2,860/W14,300 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $2/10 per MTok.
# Reasoning controls verified empirically on 2026-08-04 against the live gateway:
# OpenAI-style `reasoning_effort` and Anthropic-style `thinking` are accepted but not
# forwarded (no reasoning-token accounting, no output differential) -> affirmatively no
# caller-controllable reasoning surface; hence `reasoning_options = []`.
base_model = "anthropic/claude-sonnet-5"
reasoning_options = []

[cost]
input = 2
output = 10
19 changes: 19 additions & 0 deletions providers/cafe24/models/cafe24/auto.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,19 @@
name = "Auto Router"
description = "Automatically routes each request to a suitable model based on Cafe24 LLM Router routing policies"
family = "auto"
release_date = "2026-06-19"
last_updated = "2026-06-19"
attachment = false
reasoning = false
temperature = true
tool_call = true
structured_output = true
open_weights = false

[limit]
context = 128_000
output = 16_384

[modalities]
input = ["text"]
output = ["text"]
16 changes: 16 additions & 0 deletions providers/cafe24/models/openai/gpt-oss-120b.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
# KRW list price W56/W244 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $0.04/0.17 per MTok.
# `reasoning_effort` verified forwarded on 2026-08-04: completion tokens scale with
# effort (high=76 vs low=22 on an identical prompt).
base_model = "openai/gpt-oss-120b"

[[reasoning_options]]
type = "effort"
values = ["low", "medium", "high"]

[cost]
input = 0.04
output = 0.17

[limit]
output = 16_384
16 changes: 16 additions & 0 deletions providers/cafe24/models/openai/gpt-oss-20b.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
# KRW list price W43/W201 per MTok input/output (https://llm-router.cafe24.com/models, captured 2026-08-04).
# The site prices at a fixed W1,430/USD rate -> USD $0.03/0.14 per MTok.
# `reasoning_effort` verified forwarded on 2026-08-04: completion tokens scale with
# effort (high=76 vs low=22 on an identical prompt).
base_model = "openai/gpt-oss-20b"

[[reasoning_options]]
type = "effort"
values = ["low", "medium", "high"]

[cost]
input = 0.03
output = 0.14

[limit]
output = 16_384
5 changes: 5 additions & 0 deletions providers/cafe24/provider.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
name = "Cafe24 LLM Router"
npm = "@ai-sdk/openai-compatible"
api = "https://llm-router.cafe24.com/api/v1"
env = ["CAFE24_API_KEY"]
doc = "https://llm-router.cafe24.com/docs"
Loading