{"object": "list", "data": [{"id": "openai/whisper-large-v3-turbo", "object": "model", "created": 1780704000, "owned_by": "koscompute", "name": "OpenAI: Whisper Large v3 Turbo", "kind": "audio-transcription", "task": "audio.transcriptions", "endpoint_kind": "audio", "supported_endpoints": ["/v1/audio/transcriptions"], "input_modalities": ["audio"], "output_modalities": ["text", "transcription"], "hf_model": "openai/whisper-large-v3-turbo", "hugging_face_id": "openai/whisper-large-v3-turbo", "quantization": "int8", "context_length": 448, "modality_context_lengths": {"text": 448}, "max_output_length": 448, "max_output": 448, "pricing": {"prompt": "0", "completion": "0", "image": "0", "input_cache_read": "0", "audio_per_minute": "0.0024", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": ["temperature"], "supported_parameters": ["temperature"], "supported_features": [], "features": [], "description": "Fast speech-to-text for meeting recordings and voice notes.", "is_ready": true, "is_free": false, "openrouter": {"slug": "openai/whisper-large-v3-turbo"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 400, "capacity": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "nvidia/parakeet-tdt-0.6b-v3", "object": "model", "created": 1791936000, "owned_by": "koscompute", "name": "NVIDIA: Parakeet TDT 0.6B v3", "kind": "audio-transcription", "task": "audio.transcriptions", "endpoint_kind": "audio", "supported_endpoints": ["/v1/audio/transcriptions"], "input_modalities": ["audio"], "output_modalities": ["text", "transcription"], "hf_model": "nvidia/parakeet-tdt-0.6b-v3", "hugging_face_id": "nvidia/parakeet-tdt-0.6b-v3", "quantization": "fp32", "context_length": 0, "modality_context_lengths": {"text": 0}, "max_output_length": 0, "max_output": 0, "pricing": {"prompt": "0", "completion": "0", "image": "0", "input_cache_read": "0", "audio_per_minute": "0.0014", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": [], "supported_parameters": [], "supported_features": [], "features": [], "description": "Fast multilingual speech-to-text for 25 European languages including Czech, Slovak, German, Polish and English, with automatic punctuation, segment and word timestamps, and SRT/VTT output. Audio up to 60 minutes per request. Model by NVIDIA, licensed under CC-BY-4.0.", "is_ready": true, "is_free": false, "openrouter": {"slug": "nvidia/parakeet-tdt-0.6b-v3"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 600, "capacity": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "kosmik/tts-piper-fast", "object": "model", "created": 1750000000, "owned_by": "koscompute", "name": "Kosmik: Piper TTS Fast", "kind": "audio-speech", "task": "text-to-speech", "endpoint_kind": "tts", "supported_endpoints": ["/v1/audio/speech"], "input_modalities": ["text"], "output_modalities": ["audio"], "hf_model": "", "hugging_face_id": "", "quantization": "int8", "context_length": 0, "modality_context_lengths": {"text": 0}, "max_output_length": 0, "max_output": 0, "pricing": {"prompt": "0", "completion": "0", "image": "0", "request": "0", "input_cache_read": "0", "tts_per_1m_chars": "0.50", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": [], "supported_parameters": ["voice", "speed", "response_format"], "supported_features": [], "features": [], "description": "Fast English speech synthesis with the LJSpeech voice. WAV output. The Czech Jirka voice is unavailable pending clarification of upstream training-model licensing; use Supertonic for Czech.", "is_ready": true, "is_free": false, "openrouter": {"slug": "kosmik/tts-piper-fast"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 500, "capacity": {"available_concurrency": 16, "total_concurrency": 16, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 16, "total_concurrency": 16, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "kosmik/tts-kokoro-quality", "object": "model", "created": 1750000000, "owned_by": "koscompute", "name": "Kosmik: Kokoro TTS Quality", "kind": "audio-speech", "task": "text-to-speech", "endpoint_kind": "tts", "supported_endpoints": ["/v1/audio/speech"], "input_modalities": ["text"], "output_modalities": ["audio"], "hf_model": "", "hugging_face_id": "", "quantization": "fp32", "context_length": 0, "modality_context_lengths": {"text": 0}, "max_output_length": 0, "max_output": 0, "pricing": {"prompt": "0", "completion": "0", "image": "0", "request": "0", "input_cache_read": "0", "tts_per_1m_chars": "1.00", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": [], "supported_parameters": ["voice", "speed", "response_format"], "supported_features": [], "features": [], "description": "High-quality multilingual text-to-speech.", "is_ready": true, "is_free": false, "openrouter": {"slug": "kosmik/tts-kokoro-quality"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 200, "capacity": {"available_concurrency": 12, "total_concurrency": 12, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 12, "total_concurrency": 12, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "qwen/qwen3.8-27b", "object": "model", "created": 1786725556, "owned_by": "koscompute", "name": "Qwen: Qwen3.8 27B", "kind": "chat", "task": "chat.completions", "endpoint_kind": "chat", "supported_endpoints": ["/v1/chat/completions", "/v1/responses", "/v1/messages"], "input_modalities": ["text", "image"], "output_modalities": ["text"], "hf_model": "acyildirimer/Qwen3.8-27B-NVFP4", "hugging_face_id": "acyildirimer/Qwen3.8-27B-NVFP4", "quantization": "nvfp4", "context_length": 262144, "modality_context_lengths": {"text": 262144, "image": 262144}, "max_output_length": 32768, "max_output": 32768, "pricing": {"prompt": "0.00000035", "completion": "0.00000220", "image": "0.00000500", "request": "0", "input_cache_read": "0.00000009", "input": 0.35, "output": 2.2, "cache_read": 0.09}, "supported_sampling_parameters": ["temperature", "top_p", "top_k", "frequency_penalty", "presence_penalty", "repetition_penalty", "stop", "seed", "max_tokens", "logprobs", "top_logprobs"], "supported_parameters": ["temperature", "top_p", "top_k", "frequency_penalty", "presence_penalty", "repetition_penalty", "stop", "seed", "max_tokens", "logprobs", "top_logprobs", "include_reasoning", "reasoning", "response_format", "structured_outputs", "reasoning_effort", "enable_thinking", "preserve_thinking", "tools", "tool_choice"], "supported_features": ["tool_calling", "json_mode", "structured_outputs", "reasoning"], "features": ["tool_calling", "json_mode", "structured_outputs", "reasoning"], "description": "Qwen3.8 27B NVFP4 runs only while a Holly rental GPU is free. Rental jobs have priority; API capacity may disappear without notice.", "is_ready": true, "is_free": false, "openrouter": {"slug": "qwen/qwen3.8-27b"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 150000, "capacity": {"available_concurrency": 64, "total_concurrency": 64, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"vision": {"available_concurrency": 64, "total_concurrency": 64, "active_requests": 0, "utilization_percent": 0.0}}}, "vision": {"max_pixels": 1048576}, "architectures": [{"type": "text", "mixing": false, "topology": "2,2", "architecture": "transformer", "activation": "swiglu", "tokenizer": "Qwen2", "precision": "bfloat16", "quantization": "nvfp4", "num_parameters": {"total": 27000000000, "string": "27B"}}], "model_info": {"supported_reasoning_efforts": ["none", "low", "medium", "xhigh"], "extended_thinking": true, "reasoning": true, "thinking_default": true, "max_thinking_tokens": 8192}}, {"id": "kosmik/tts-moss-nano", "object": "model", "created": 1750000000, "owned_by": "koscompute", "name": "Kosmik: MOSS-TTS-Nano", "kind": "audio-speech", "task": "text-to-speech", "endpoint_kind": "tts", "supported_endpoints": ["/v1/audio/speech"], "input_modalities": ["text"], "output_modalities": ["audio"], "hf_model": "OpenMOSS-Team/MOSS-TTS-Nano-100M-ONNX", "hugging_face_id": "OpenMOSS-Team/MOSS-TTS-Nano-100M-ONNX", "quantization": "fp32", "context_length": 0, "modality_context_lengths": {"text": 0}, "max_output_length": 0, "max_output": 0, "pricing": {"prompt": "0", "completion": "0", "image": "0", "request": "0", "input_cache_read": "0", "tts_per_1m_chars": "4.00", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": [], "supported_parameters": ["voice", "response_format"], "supported_features": [], "features": [], "description": "Multilingual speech synthesis with built-in voices, including Czech. WAV output; speed must be 1. Custom reference voices are not supported by this endpoint.", "is_ready": true, "is_free": false, "openrouter": {"slug": "kosmik/tts-moss-nano"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 40, "capacity": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 2, "total_concurrency": 2, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "kosmik/tts-supertonic-3", "object": "model", "created": 1750000000, "owned_by": "koscompute", "name": "Kosmik: Supertonic 3", "kind": "audio-speech", "task": "text-to-speech", "endpoint_kind": "tts", "supported_endpoints": ["/v1/audio/speech"], "input_modalities": ["text"], "output_modalities": ["audio"], "hf_model": "Supertone/supertonic-3", "hugging_face_id": "Supertone/supertonic-3", "quantization": "fp32", "context_length": 0, "modality_context_lengths": {"text": 0}, "max_output_length": 0, "max_output": 0, "pricing": {"prompt": "0", "completion": "0", "image": "0", "request": "0", "input_cache_read": "0", "tts_per_1m_chars": "0.80", "input": 0.0, "output": 0.0, "cache_read": 0.0}, "supported_sampling_parameters": [], "supported_parameters": ["voice", "speed", "response_format"], "supported_features": [], "features": [], "description": "Multilingual speech synthesis. WAV output, speed 0.5\u20132. Use a language-prefixed voice such as en:M1 or cs:F1; unprefixed voices default to Czech. Quality uses 8 steps. Use is subject to the Open RAIL-M license and restrictions: https://api.koscompute.com/docs/security/speech-model-licenses/", "is_ready": true, "is_free": false, "openrouter": {"slug": "kosmik/tts-supertonic-3"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 300, "capacity": {"available_concurrency": 4, "total_concurrency": 4, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 4, "total_concurrency": 4, "active_requests": 0, "utilization_percent": 0.0}}}}, {"id": "kosmik/mimo-v2.6-flash-uncensored", "object": "model", "created": 1791503000, "owned_by": "koscompute", "name": "MiMo-V2.6 Flash RL Uncensored (Heretic)", "kind": "chat", "task": "chat.completions", "endpoint_kind": "chat", "supported_endpoints": ["/v1/chat/completions", "/v1/responses", "/v1/messages"], "input_modalities": ["text"], "output_modalities": ["text"], "hf_model": "MorinoNushi/MiMo-V2.6-Flash-RL-Uncensored-Heretic-GGUF", "hugging_face_id": "XiaomiMiMo/MiMo-V2.6-Flash-RL", "quantization": "gguf-iq2-mixed", "context_length": 262144, "modality_context_lengths": {"text": 262144}, "max_output_length": 32768, "max_output": 32768, "pricing": {"prompt": "0.00000015", "completion": "0.00000045", "image": "0", "request": "0", "input_cache_read": "0.00000003", "input": 0.15, "output": 0.45, "cache_read": 0.03}, "supported_sampling_parameters": ["temperature", "top_p", "top_k", "frequency_penalty", "presence_penalty", "repetition_penalty", "stop", "seed", "max_tokens"], "supported_parameters": ["temperature", "top_p", "top_k", "frequency_penalty", "presence_penalty", "repetition_penalty", "stop", "seed", "max_tokens", "include_reasoning", "reasoning", "response_format", "structured_outputs", "reasoning_effort", "enable_thinking", "tools", "tool_choice"], "supported_features": ["tool_calling", "json_mode", "structured_outputs", "reasoning"], "features": ["tool_calling", "json_mode", "structured_outputs", "reasoning"], "description": "Uncensored MiMo-V2.6 Flash RL (Heretic refusal-removal adapter) for text chat, reasoning and tools. Runs only on rental spare capacity: rentals take priority and the model can become unavailable without notice; handle 503 with backoff. Reduced guardrails: you are responsible for how outputs are used.", "is_ready": true, "is_free": false, "openrouter": {"slug": "kosmik/mimo-v2.6-flash-uncensored"}, "datacenters": [{"country_code": "CZ"}], "capacity_tpm": 7000, "capacity": {"available_concurrency": 16, "total_concurrency": 16, "active_requests": 0, "utilization_percent": 0.0, "unit": "concurrent_requests", "updated_at": 1792636681, "modalities": {"text": {"available_concurrency": 16, "total_concurrency": 16, "active_requests": 0, "utilization_percent": 0.0}}}, "model_info": {"supported_reasoning_efforts": ["none", "low", "medium", "high"]}}]}