{"data":[{"schema_version":"2.4","id":"glm-5.3-nvfp4","name":"GLM-5.3 NVFP4","created":1787616000,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.000001206"},{"type":"cached_prompt","unit":"token","cost_usd":"0.000000224"}]}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.000003789"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"RadixArk/GLM-5.3-NVFP4","quantization":"nvfp4","description":"GLM-5.3 753B (40B active MoE) NVFP4 on 8x NVIDIA B200 in Madrid. 1M context. Reasoning is always on — set effort to low (default) or high; it cannot be disabled. Tool calling supported.","openrouter":{"slug":"z-ai/glm-5.3"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"qwen-3-14b","name":"Qwen3 14B FP8","created":1745884800,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":32768,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000002"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000002"}]}],"output_modalities":[{"type":"text","max_length":{"value":8192,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":8192,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.000001"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"Qwen/Qwen3-14B-FP8","quantization":"fp8","description":"Qwen3 14B FP8 reasoning model on MIG-sliced B200 in Madrid. Supports thinking mode with configurable budget.","openrouter":{"slug":"qwen/qwen3-14b"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"qwen-3.5-9b","name":"Qwen3.5 9B","created":1772409600,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":65536,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000001"}]}],"output_modalities":[{"type":"text","max_length":{"value":8192,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":8192,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000015"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"Qwen/Qwen3.5-9B","quantization":"fp8","description":"Qwen3.5 9B lightweight reasoning model on MIG-sliced B200 in Madrid. Fast and cost-effective. Supports tool calling.","openrouter":{"slug":"qwen/qwen3.5-9b"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"qwen-3.8-27b","name":"Qwen3.8 27B","created":1786752000,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":98304,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000001"}]}],"output_modalities":[{"type":"text","max_length":{"value":8192,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":8192,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000003"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"Inferact/Qwen3.8-27B-NVFP4","quantization":"nvfp4","description":"Qwen3.8 27B (NVFP4) on a MIG-sliced B200 in Madrid. Strong coding and agentic capabilities. Supports tool calling. Replaces Qwen3.6 27B.","openrouter":{"slug":"qwen/qwen3.8-27b"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}],"capacity":[{"type":"concurrency","unit":"request","value":8}]},{"schema_version":"2.4","id":"gemma-4-26b-a4b","name":"Gemma 4 26B A4B IT","created":1777507200,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":32768,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000012"},{"type":"cached_prompt","unit":"token","cost_usd":"0.000000012"}]},{"type":"image","supported_inputs":{"sources":{"type":"enum","values":["url","base64"]},"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]},"references":{"type":"integer","min":0,"max":4}}}],"output_modalities":[{"type":"text","max_length":{"value":8192,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":8192,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000003"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"nvidia/Gemma-4-26B-A4B-NVFP4","quantization":"nvfp4","description":"Gemma 4 26B-A4B MoE (~3.8B active) in NVFP4 on MIG-sliced B200 in Madrid. Multimodal (text+image) with reasoning and tool calling. Replaces Gemma 4 E2B/E4B.","openrouter":{"slug":"google/gemma-4-26b-a4b-it"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"nemotron-nano-12b-v2-vl","name":"Nemotron Nano 12B v2 VL","created":1761609600,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":131072,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000006"}]},{"type":"image","supported_inputs":{"sources":{"type":"enum","values":["url","base64"]},"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]},"references":{"type":"integer","min":0,"max":4}}}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000024"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-NVFP4-QAD","quantization":"nvfp4","description":"NVIDIA Nemotron Nano 12B v2 VL (NVFP4) on MIG-sliced B200 in Madrid. Vision-language for OCR and document understanding — 128K context, up to 4 images per prompt.","datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"nemotron-3.5-lightning-30b-a3b","name":"Nemotron 3.5 Lightning 30B A3B","created":1786406400,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":131072,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000012"}]}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"reasoning":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000003"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4","quantization":"nvfp4","description":"NVIDIA Nemotron 3.5 Lightning 30B-A3B MoE (~3B active, Mamba-2 hybrid) in NVFP4 on MIG-sliced B200 in Madrid. Reasoning and tool calling, 128K context.","openrouter":{"slug":"nvidia/nemotron-3.5-lightning"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"qwen3-vl-32b-ocr","name":"Qwen3-VL 32B OCR","created":1761004800,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":65536,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001"}]},{"type":"image","supported_inputs":{"sources":{"type":"enum","values":["url","base64"]},"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]},"references":{"type":"integer","min":0,"max":4}}}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000035"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"RedHatAI/Qwen3-VL-32B-Instruct-NVFP4","quantization":"nvfp4","description":"Qwen3-VL 32B Instruct (NVFP4) on a MIG-sliced B200 in Madrid. Vision-language for OCR and document understanding — 64K context, up to 4 images per prompt.","openrouter":{"slug":"qwen/qwen3-vl-32b-instruct"},"datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"alia-40b-instruct","name":"ALIA 40B Instruct","created":1782777600,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":163840,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000002"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000002"}]}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"},"structured_outputs":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.000002"}]}],"is_ready":true,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"BSC-LT/ALIA-40b-instruct-2606","quantization":"bf16","description":"ALIA 40B Instruct (BSC, Spain's public LLM) in bf16 on 1x NVIDIA B200 in Madrid. Dense 40B, 160K context; trained for Spanish, Catalan, Basque, Galician and English. Chat and JSON mode; no thinking mode. BSC recommends temperature 0-0.2 without repetition penalty.","datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"llama-3.1-70b-instruct","name":"Llama 3.1 70B Instruct","created":1788861813,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":65536,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000035"},{"type":"cached_prompt","unit":"token","cost_usd":"0.000000035"}]}],"output_modalities":[{"type":"text","max_length":{"value":16384,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":16384,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000014"}]}],"is_ready":false,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"meta-llama/Meta-Llama-3.1-70B-Instruct","quantization":"bf16","description":"Llama 3.1 70B Instruct in bf16 on NVIDIA B200 in Madrid. Dense 70B, 64K context as served. Tool calling and JSON mode; no thinking mode.","datacenters":[{"country_code":"ES","region":"es-mad-1"}]},{"schema_version":"2.4","id":"llama-3.1-8b-instruct","name":"Llama 3.1 8B Instruct","created":1788861813,"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":8192,"unit":"token"}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000001"}]}],"output_modalities":[{"type":"text","max_length":{"value":4096,"unit":"token"},"streaming":true,"supported_parameters":{"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"top_k":{"type":"integer","min":-1,"max":1024},"min_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"repetition_penalty":{"type":"range","min":0,"max":2},"stop":{"type":"array","max_items":16},"seed":{"type":"integer","min":0,"max":2147483647},"max_tokens":{"type":"integer","min":1,"max":4096,"unit":"token"},"logit_bias":{"type":"unknown"},"tools":{"type":"boolean"},"structured_outputs":{"type":"boolean"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000003"}]}],"is_ready":false,"compliance":{"zdr":true,"hipaa":false},"hugging_face_id":"meta-llama/Llama-3.1-8B-Instruct","quantization":"bf16","description":"Llama 3.1 8B Instruct in bf16 on NVIDIA B200 in Madrid. Dense 8B, 8K context as served — the engine is capped well below the model's nominal window. Tool calling and JSON mode.","datacenters":[{"country_code":"ES","region":"es-mad-1"}]}]}