{"data":[{"schema_version":"2.4","id":"qwen38","name":"Qwen3.8 27B","hugging_face_id":"Qwen/Qwen3.8-27B","description":"Qwen3.8 27B in NVFP4 with the multi-token prediction head, served on dedicated GPUs in the European Economic Area. Full 262k context. Prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":true,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"qwen/qwen3.8-27b"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":262144}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000010000"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000000761"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000010000"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000030000"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"qwen38-unc","name":"Qwen3.8 27B Uncensored","hugging_face_id":"huihui-ai/Huihui-Qwen3.8-27B-abliterated","description":"Qwen3.8 27B Uncensored, NVFP4 build with the multi-token prediction head, served on dedicated GPUs in the European Economic Area. Context up to 1M tokens, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"qwen/qwen3.8-27b-uncensored"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004940"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000000703"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":32000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004940"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":235929},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000021278"}]}],"quantization":"fp8"},{"schema_version":"2.4","id":"k2-horizon-7b","name":"K2-Horizon 7B","hugging_face_id":"IFM/K2-Horizon-7B","description":"K2-Horizon 7B. Currently offline: the GPU has been released.","tokenizer":"Other","is_ready":false,"deployment_region":"US","compliance":{"zdr":true},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":131072}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000040"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000030"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":1500000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":16384},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000120"}]}],"quantization":"fp8"},{"schema_version":"2.4","id":"qwen38-obl","name":"Qwen3.8 27B Obliterated","hugging_face_id":"OBLITERATUS/Qwen3.8-27B-OBLITERATED","description":"Qwen3.8 27B Obliterated served as an NVFP4 Qwen3.8-27B base with the variant's weight delta applied as a LoRA adapter on dedicated GPUs in the European Economic Area. Full 512k context, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"qwen/qwen3.8-27b-obliterated"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000002000"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000001800"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000002000"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000014000"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"qwen38-queen","name":"Qwen3.8 27B Queen","hugging_face_id":"aifeifei798/Qwen3.8-Queen-27B","description":"Qwen3.8 27B Queen served as an NVFP4 Qwen3.8-27B base with the variant's weight delta applied as a LoRA adapter on dedicated GPUs in the European Economic Area. Full 512k context, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":true,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"qwen/qwen3.8-27b-queen"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000000950"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000000448"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000000950"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000005800"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"qwen38-fable","name":"Qwen3.8 27B Fable","hugging_face_id":"TeichAI/Qwen3.8-27B-Fable-Distill","description":"Qwen3.8 27B Fable served as an NVFP4 Qwen3.8-27B base with the variant's weight delta applied as a LoRA adapter on dedicated GPUs in the European Economic Area. Full 512k context, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":true,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"qwen/qwen3.8-27b-fable"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000000950"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000000448"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000000950"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000005800"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"heabsy-v1","name":"Heabsy V1","description":"Heabsy V1 is our own assistant model, fine-tuned by Heabsy AI on an open-weight foundation. It speaks 30+ languages natively, returns its reasoning separately with adjustable effort (high / medium / low / off), calls tools reliably, and handles a full 524k-token context. Served on dedicated GPUs in the European Economic Area; prompts and completions are never stored.","tokenizer":"Other","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"heabsy/heabsy-v1"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000009523809524"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000002857142857"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000028571428571"}]}],"quantization":"fp8"},{"schema_version":"2.4","id":"cyberheabsy-v1-small","name":"CyberHeabsy V1 Small","description":"CyberHeabsy V1 is an unrestricted assistant by Heabsy AI, fine-tuned on an open-weight foundation for security research, red-teaming, uncensored roleplay and creative writing with minimal refusals. Full separate reasoning with adjustable effort, native tool calling and a 524k-token context. Served on dedicated GPUs; prompts and completions are never stored.","tokenizer":"Other","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"heabsy/cyberheabsy-v1-small"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000009523809524"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000002857142857"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":4000000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":32768},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000028571428571"}]}],"quantization":"fp8"},{"schema_version":"2.4","id":"mimo-v26-flash-unc","name":"MiMo V2.6 Flash Uncensored","hugging_face_id":"dealignai/MiMo-V2.6-Flash-RL-UNCENSORED","description":"MiMo-V2.6-Flash-MOPD Uncensored (dealignai weight-level refusal removal carried over to the MOPD weights), NVFP4 experts with the DFlash speculative-decoding head, served on dedicated GPUs. 1M context, vision input, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":true,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"xiaomi/mimo-v2.6-flash-uncensored"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004404"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000001404"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004404"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000014285"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"mimo-v26-flash","name":"MiMo V2.6 Flash","hugging_face_id":"XiaomiMiMo/MiMo-V2.6-Flash-MOPD","description":"MiMo-V2.6-Flash-MOPD (official Xiaomi weights; the MOPD update of Flash-RL with fewer repeated tool calls in agent loops), NVFP4 experts with the DFlash speculative-decoding head, served on dedicated GPUs. 1M context, vision input, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":true,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"xiaomi/mimo-v2.6-flash"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004762"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000001900"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004762"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000014286"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"mimo-v26-flash-abl","name":"MiMo V2.6 Flash Abliterated","hugging_face_id":"MorinoNushi/MiMo-V2.6-Flash-RL-Uncensored-Heretic-LoRA-GGUF","description":"MiMo-V2.6-Flash-MOPD Abliterated: the refusal direction (Heretic directional ablation) removed from the attention output projections, applied as a LoRA adapter on an NVFP4 base with the DFlash speculative-decoding head, served on dedicated GPUs. 1M context, vision input, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":true,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"xiaomi/mimo-v2.6-flash-abliterated"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004762"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000002571"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000004762"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000014286"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"cyberheabsy","name":"CyberHeabsy","hugging_face_id":"heabsy/CyberHeabsy","description":"CyberHeabsy is a high-performance language model purpose-built for cybersecurity, code analysis, and vulnerability research.\n\nKey Features:\n\nCybersecurity Focus: Tailored specifically for red-teaming, security auditing, reverse engineering, and advanced malware analysis.\n\nDeep Analytical Reasoning: Employs integrated, step-by-step cognitive processing to solve complex, multi-stage infrastructure challenges.\n\nTarget Audience: Security researchers, penetration testers, security engineers, and developers requiring a secure, strictly private AI assistant for advanced threat analysis.","tokenizer":"Other","is_ready":true,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"heabsy/cyberheabsy"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000014286"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000002571"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000042857"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"deepseek-v4.1-flash","name":"DeepSeek V4.1 Flash","hugging_face_id":"deepseek-ai/DeepSeek-V4.1-Flash","description":"DeepSeek-V4.1-Flash from DeepSeek's released weights: routed experts in their native MXFP4 format, attention and dense layers FP8, Engram tables int4. During decoding a small share of low-weight expert routes is skipped for speed; measured quality matches the full model on GSM8K and HumanEval. Served on dedicated GPUs in the European Economic Area (Finland). 1M context, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"deepseek/deepseek-v4.1-flash"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000011428571"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000000285714"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000042857143"}]}]},{"schema_version":"2.4","id":"deepseek-v4.1-flash-unc","name":"DeepSeek V4.1 Flash Uncensored","hugging_face_id":"diffbot/DeepSeek-V4.1-Flash-EXL3-2.0bpw-2x-RTX-PRO-6000","description":"DeepSeek-V4.1-Flash Uncensored: refusal removal from dealignai's DeepSeek-V4.1-Flash-UNCENSORED applied as a LoRA delta on attention output and shared-expert projections; routed experts compressed to EXL3 2.0 bpw (attention, shared experts and Engram tables stay FP8), served on dedicated GPUs in the European Economic Area. 1M context, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":false,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"deepseek/deepseek-v4.1-flash-uncensored"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.00000047619048"},{"type":"cached_prompt","unit":"token","cost_usd":"0.00000025714286"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":2000000}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":65536},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.00000142857143"}]}]},{"schema_version":"2.4","id":"glm53-flash-unc","name":"GLM 5.3 Flash Uncensored","hugging_face_id":"dealignai/GLM-5.3-Flash-UNCENSORED-FP8","description":"GLM-5.3-Flash Uncensored (dealignai weight-level refusal removal baked into the weights, no adapter), NVFP4 experts with the multi-token prediction head, served on dedicated GPUs. Full 1M context window, vision input, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":true,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"z-ai/glm-5.3-flash-uncensored"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001739"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000000557"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":3000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000001739"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":131072},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000006957"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"qwen38-unc-fast","name":"Qwen3.8 27B Uncensored Fast by Heabsy","hugging_face_id":"heabsy/Qwen3.8-27B-Uncensored-Fast","description":"Heabsy's own build of Qwen3.8 27B Uncensored: weights we quantized and calibrated ourselves, served on our inference stack. Refusal behaviour is removed, and quality matches the original Qwen3.8 27B (benchmarks on the model card). Reasoning comes back in a separate field, tool calling works, context is 524k tokens. Served only by Heabsy, in the European Economic Area; prompts and completions are not stored in request logs.","tokenizer":"Qwen","is_ready":true,"deployment_region":"EU","compliance":{"zdr":true},"openrouter":{"slug":"heabsy/qwen3.8-27b-uncensored-fast"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":524288}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000007904"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000001429"}],"capacity":[{"type":"prompt","unit":"token","per":"minute","value":32000000}]},{"type":"image","supported_inputs":{"formats":{"type":"enum","values":["image/png","image/jpeg","image/webp"]}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000007904"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":235929},"temperature":{"type":"range","min":0,"max":2},"top_p":{"type":"range","min":0,"max":1},"frequency_penalty":{"type":"range","min":-2,"max":2},"presence_penalty":{"type":"range","min":-2,"max":2},"stop":{"type":"array"},"seed":{"type":"integer"}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000024952"}]}],"quantization":"nvfp4"},{"schema_version":"2.4","id":"glm53","name":"GLM 5.3","description":"GLM-5.3 served by Heabsy on dedicated GPUs. Context up to 1M tokens, reasoning returned separately, tool calling enabled. Prompts and completions are not stored in request logs.","tokenizer":"Other","is_ready":false,"deployment_region":"global","compliance":{"zdr":true},"openrouter":{"slug":"z-ai/glm-5.3"},"input_modalities":[{"type":"text","supported_inputs":{"max_context_length":{"value":1048576}},"pricing":[{"type":"prompt","unit":"token","cost_usd":"0.0000005217"},{"type":"cached_prompt","unit":"token","cost_usd":"0.0000001739"}]}],"output_modalities":[{"type":"text","streaming":true,"supported_parameters":{"max_tokens":{"type":"integer","min":1,"max":131072}},"pricing":[{"type":"completion","unit":"token","cost_usd":"0.0000009565"}]}],"quantization":"nvfp4"}]}