diff --git a/models/deepseek/deepseek-ocr-2.toml b/models/deepseek/deepseek-ocr-2.toml new file mode 100644 index 0000000000..9e55df60d6 --- /dev/null +++ b/models/deepseek/deepseek-ocr-2.toml @@ -0,0 +1,16 @@ +name = "DeepSeek OCR 2" +description = "High-accuracy OCR model for extracting text from documents, screenshots, receipts, and natural scenes" +release_date = "2026-01-27" +last_updated = "2026-01-27" +attachment = true +reasoning = false +tool_call = false +open_weights = true + +[limit] +context = 8_192 +output = 8_192 + +[modalities] +input = ["text", "image"] +output = ["text"] diff --git a/models/swiss-ai/apertus-70b.toml b/models/swiss-ai/apertus-70b.toml new file mode 100644 index 0000000000..483e6ccdbf --- /dev/null +++ b/models/swiss-ai/apertus-70b.toml @@ -0,0 +1,28 @@ +name = "Apertus 70B" +description = "Fully open 70B multilingual LLM supporting 1800+ languages with 65K context. Trained on 15T tokens of compliant open data. Apache 2.0, EU AI Act compliant." +release_date = "2025-09-02" +last_updated = "2025-09-02" +knowledge = "2025-09" +attachment = false +reasoning = false +temperature = true +tool_call = true +open_weights = true +license = "Apache-2.0" + +[limit] +context = 65_536 +output = 8_192 + +[modalities] +input = ["text"] +output = ["text"] + +[[weights]] +label = "Hugging Face" +url = "https://huggingface.co/swiss-ai/Apertus-70B-Instruct-2509" + +[[links]] +label = "Paper" +url = "https://arxiv.org/abs/2509.14233" +type = "paper" diff --git a/providers/regolo-ai/models/apertus-70b.toml b/providers/regolo-ai/models/apertus-70b.toml new file mode 100644 index 0000000000..76656378c3 --- /dev/null +++ b/providers/regolo-ai/models/apertus-70b.toml @@ -0,0 +1,10 @@ +# Prices listed in EUR on Regolo; converted to USD at 1 EUR = 1.1537 USD (2026-08-05) +base_model = "swiss-ai/apertus-70b" + +[cost] +input = 0.46 +output = 2.42 + +[limit] +context = 30_000 +output = 30_000 diff --git a/providers/regolo-ai/models/brick-complexity-pro.toml b/providers/regolo-ai/models/brick-complexity-pro.toml new file mode 100644 index 0000000000..5776215d8c --- /dev/null +++ b/providers/regolo-ai/models/brick-complexity-pro.toml @@ -0,0 +1,22 @@ +# Prices listed in EUR on Regolo; converted to USD at 1 EUR = 1.1537 USD (2026-08-05) +name = "Brick Complexity Pro" +description = "Complexity classifier that powers the Brick semantic router by extracting query difficulty" +release_date = "2026-02-06" +last_updated = "2026-02-06" +attachment = true +reasoning = false +temperature = true +tool_call = true +open_weights = false + +[cost] +input = 0.12 +output = 0.46 + +[limit] +context = 100_000 +output = 15_000 + +[modalities] +input = ["text", "image"] +output = ["text"] diff --git a/providers/regolo-ai/models/brick-v1-beta.toml b/providers/regolo-ai/models/brick-v1-beta.toml new file mode 100644 index 0000000000..9f83280c1e --- /dev/null +++ b/providers/regolo-ai/models/brick-v1-beta.toml @@ -0,0 +1,23 @@ +name = "Brick v1 Beta" +description = "Semantic router by Regolo.ai that directs each request to the most suitable model, optimizing costs and performance" +family = "model-router" +release_date = "2026-02-06" +last_updated = "2026-02-06" +attachment = true +reasoning = false +temperature = true +tool_call = true +open_weights = false +status = "beta" + +[cost] +input = 0.00 +output = 0.00 + +[limit] +context = 100_000 +output = 15_000 + +[modalities] +input = ["text", "image"] +output = ["text"] diff --git a/providers/regolo-ai/models/deepseek-ocr-2.toml b/providers/regolo-ai/models/deepseek-ocr-2.toml new file mode 100644 index 0000000000..0b598c83bb --- /dev/null +++ b/providers/regolo-ai/models/deepseek-ocr-2.toml @@ -0,0 +1,10 @@ +# Free on Regolo (input_cost_per_token = 0, output_cost_per_token = 0 per API 2026-08-05) +base_model = "deepseek/deepseek-ocr-2" + +[cost] +input = 0.00 +output = 0.00 + +[limit] +context = 4_000 +output = 4_000 diff --git a/providers/regolo-ai/models/faster-whisper-large-v3.toml b/providers/regolo-ai/models/faster-whisper-large-v3.toml new file mode 100644 index 0000000000..3ff4d8dbb7 --- /dev/null +++ b/providers/regolo-ai/models/faster-whisper-large-v3.toml @@ -0,0 +1,8 @@ +# Free on Regolo (input_cost_per_token = 0, output_cost_per_token = 0 per API 2026-08-05) +# Faster-Whisper (CTranslate2) build of Whisper Large v3, not the OpenAI reference model +base_model = "openai/whisper-large-v3" +name = "Faster Whisper Large v3" + +[cost] +input = 0.00 +output = 0.00 diff --git a/providers/regolo-ai/models/gemma4-31b.toml b/providers/regolo-ai/models/gemma4-31b.toml new file mode 100644 index 0000000000..9c167710fb --- /dev/null +++ b/providers/regolo-ai/models/gemma4-31b.toml @@ -0,0 +1,13 @@ +# Prices listed in EUR on Regolo; converted to USD at 1 EUR = 1.1537 USD (2026-08-05) +# Toggle: thinking = true|false +base_model = "google/gemma-4-31b-it" + +reasoning_options = [{ type = "toggle" }] + +[cost] +input = 0.46 +output = 2.42 + +[limit] +context = 100_000 +output = 100_000 diff --git a/providers/regolo-ai/models/glm5.2.toml b/providers/regolo-ai/models/glm5.2.toml new file mode 100644 index 0000000000..2882d657ba --- /dev/null +++ b/providers/regolo-ai/models/glm5.2.toml @@ -0,0 +1,14 @@ +# Prices listed in EUR on Regolo; converted to USD at 1 EUR = 1.1537 USD (2026-08-05) +# Effort: reasoning_effort = high|max (lab maps low/medium→high, xhigh→max; none skips thinking) +# Source: https://docs.bigmodel.cn/cn/guide/capabilities/thinking (accessed 2026-06-25) +base_model = "zhipuai/glm-5.2" + +reasoning_options = [{ type = "effort", values = ["high", "max"] }] + +[cost] +input = 2.31 +output = 6.00 + +[limit] +context = 96_000 +output = 96_000 diff --git a/providers/regolo-ai/models/llama-3.1-8b-instruct.toml b/providers/regolo-ai/models/llama-3.1-8b-instruct.toml deleted file mode 100644 index c515060b64..0000000000 --- a/providers/regolo-ai/models/llama-3.1-8b-instruct.toml +++ /dev/null @@ -1,22 +0,0 @@ -name = "Llama 3.1 8B Instruct" -description = "Open Llama instruction model for multilingual chat, reasoning, and coding" -family = "llama" -release_date = "2025-04-07" -last_updated = "2025-04-07" -attachment = false -reasoning = false -temperature = true -tool_call = true -open_weights = false - -[cost] -input = 0.05 -output = 0.25 - -[limit] -context = 120000 -output = 120000 - -[modalities] -input = ["text"] -output = ["text"] \ No newline at end of file diff --git a/providers/regolo-ai/models/minimax-m2.5.toml b/providers/regolo-ai/models/minimax-m2.5.toml deleted file mode 100644 index 8126a5b23d..0000000000 --- a/providers/regolo-ai/models/minimax-m2.5.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "MiniMax 2.5" -description = "MiniMax model for chat, coding, office work, and agentic tasks" -family = "minimax" -release_date = "2026-03-10" -last_updated = "2026-03-10" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -open_weights = false - -[cost] -input = 0.80 -output = 3.50 - -[limit] -context = 190000 -output = 64000 - -[modalities] -input = ["text"] -output = ["text"] \ No newline at end of file diff --git a/providers/regolo-ai/models/mistral-small3.2.toml b/providers/regolo-ai/models/mistral-small3.2.toml deleted file mode 100644 index fccd72cac0..0000000000 --- a/providers/regolo-ai/models/mistral-small3.2.toml +++ /dev/null @@ -1,23 +0,0 @@ -name = "Mistral Small 3.2" -description = "Efficient Mistral model for fast chat, extraction, and production assistants" -family = "mistral-small" -release_date = "2025-01-31" -last_updated = "2025-01-31" -attachment = false -reasoning = true -reasoning_options = [] -temperature = true -tool_call = true -open_weights = false - -[cost] -input = 0.50 -output = 2.20 - -[limit] -context = 120000 -output = 120000 - -[modalities] -input = ["text"] -output = ["text"] \ No newline at end of file diff --git a/providers/regolo-ai/models/qwen3.6-27b.toml b/providers/regolo-ai/models/qwen3.6-27b.toml new file mode 100644 index 0000000000..e10771de9f --- /dev/null +++ b/providers/regolo-ai/models/qwen3.6-27b.toml @@ -0,0 +1,17 @@ +# Prices listed in EUR on Regolo; converted to USD at 1 EUR = 1.1537 USD (2026-08-05) +# Toggle: thinking = true|false +base_model = "alibaba/qwen3.6-27b" + +reasoning_options = [{ type = "toggle" }] + +[cost] +input = 0.58 +output = 2.42 + +[limit] +context = 120_000 +output = 120_000 + +[modalities] +input = ["text", "image", "video"] +output = ["text"]