{"models":[{"id":"claude-haiku-4-5","label":"Claude Haiku 4.5","description":"Summarization, classification, lightweight tool use.","task_type":"decoder","context_window":200000,"input_price_per_million":1.0,"output_price_per_million":5.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"claude-opus-4-7","label":"Claude Opus 4.7","description":"Complex reasoning, long documents, nuanced writing.","task_type":"decoder","context_window":1000000,"input_price_per_million":5.0,"output_price_per_million":25.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"claude-sonnet-4-6","label":"Claude Sonnet 4.6","description":"Coding, writing, multi-step reasoning, tool use.","task_type":"decoder","context_window":1000000,"input_price_per_million":3.0,"output_price_per_million":15.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"deepseek-ai/DeepSeek-V3.1","label":"DeepSeek V3.1","description":"Coding, math, long-form reasoning.","task_type":"decoder","context_window":163840,"input_price_per_million":0.56,"output_price_per_million":1.68,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"MIT","tier":"open"},{"id":"deepseek-ai/DeepSeek-V4-Flash","label":"DeepSeek V4 Flash","description":"Fast, cost-efficient long-context reasoning and generation.","task_type":"decoder","context_window":1000000,"input_price_per_million":0.14,"output_price_per_million":0.28,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"MIT","tier":"open"},{"id":"deepseek-ai/DeepSeek-V4-Pro","label":"DeepSeek V4 Pro","description":"Coding, math, extended chain-of-thought reasoning.","task_type":"decoder","context_window":1000000,"input_price_per_million":1.7,"output_price_per_million":3.4,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"MIT","tier":"open"},{"id":"fastino/gliguard-LLMGuardrails-300M","label":"GLiGuard LLM Guardrails 300M","description":"LLM safety, PII, and adversarial-attack guardrails; 300M-param encoder.","task_type":"encoder","context_window":8192,"input_price_per_million":0.25,"output_price_per_million":0.25,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Apache-2.0","tier":"open"},{"id":"fastino/gliner2-base-v1","label":"GLiNER2 Base","description":"Named entity recognition; zero-shot span extraction.","task_type":"encoder","context_window":8192,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"fastino/gliner2-large-v1","label":"GLiNER2 Large","description":"Higher-accuracy NER; broader entity coverage.","task_type":"encoder","context_window":8192,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"fastino/gliner2-multi-v1","label":"GLiNER2 Multi","description":"Multilingual NER for cross-lingual extraction.","task_type":"encoder","context_window":8192,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"fastino/gliner2-multi-large-v1","label":"GLiNER2 Multi Large","description":"High-accuracy multilingual NER and span extraction.","task_type":"encoder","context_window":8192,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"fastino/gliner2-privacy-filter-PII-multi","label":"GLiNER2 Privacy Filter PII (Multi)","description":"Multilingual PII entity detection across 40+ languages; non-trainable encoder.","task_type":"encoder","context_window":8192,"input_price_per_million":0.25,"output_price_per_million":0.25,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Apache-2.0","tier":"open"},{"id":"zai-org/GLM-5.1","label":"GLM 5.1","description":"Bilingual Chinese-English reasoning, coding, and long-context tasks.","task_type":"decoder","context_window":202752,"input_price_per_million":1.05,"output_price_per_million":3.5,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"MIT","tier":"open"},{"id":"google/gemma-3-4b-pt","label":"Gemma 3 4B (Pretrained)","description":"Gemma 3 4B pretrained base — PostTrainBench post-training target.","task_type":"decoder","context_window":32768,"input_price_per_million":0.15,"output_price_per_million":0.15,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Gemma","tier":"open"},{"id":"meta-llama/Llama-3.1-8B-Instruct","label":"Llama 3.1 8B Instruct","description":"Fast classification, structured extraction, short-form generation.","task_type":"decoder","context_window":16384,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Llama-3-Community","tier":"open"},{"id":"meta-llama/Llama-3.2-1B","label":"Llama 3.2 1B","description":"Ultra-compact base; latency-critical fine-tuning.","task_type":"decoder","context_window":131072,"input_price_per_million":0.1,"output_price_per_million":0.1,"supports_inference":false,"is_chat_model":false,"supports_on_demand_inference":true,"supports_training":false,"license":"Llama-3-Community","tier":"open"},{"id":"meta-llama/Llama-3.2-1B-Instruct","label":"Llama 3.2 1B Instruct","description":"Lightweight instruction-following for edge deployment.","task_type":"decoder","context_window":60000,"input_price_per_million":0.1,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Llama-3-Community","tier":"open"},{"id":"meta-llama/Llama-3.2-3B","label":"Llama 3.2 3B","description":"Small base model for fine-tuning and fast generation.","task_type":"decoder","context_window":131072,"input_price_per_million":0.1,"output_price_per_million":0.1,"supports_inference":false,"is_chat_model":false,"supports_on_demand_inference":false,"supports_training":false,"license":"Llama-3-Community","tier":"open"},{"id":"meta-llama/Llama-3.2-3B-Instruct","label":"Llama 3.2 3B Instruct","description":"Compact chat model for low-latency deployment.","task_type":"decoder","context_window":80000,"input_price_per_million":0.1,"output_price_per_million":0.34,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Llama-3-Community","tier":"open"},{"id":"meta-llama/Llama-3.3-70B-Instruct","label":"Llama 3.3 70B Instruct","description":"RAG, complex reasoning, multi-step agentic tool use.","task_type":"decoder","context_window":131072,"input_price_per_million":0.9,"output_price_per_million":0.9,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Llama-3-Community","tier":"open"},{"id":"MiniMaxAI/MiniMax-M2.7","label":"MiniMax M2.7","description":"Long-context generation; strong on complex document tasks.","task_type":"decoder","context_window":196608,"input_price_per_million":0.3,"output_price_per_million":1.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"MIT","tier":"open"},{"id":"moonshotai/Kimi-K2.6","label":"Kimi K2.6","description":"Long-context reasoning and agentic tool use.","task_type":"decoder","context_window":256000,"input_price_per_million":0.95,"output_price_per_million":4.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Modified-MIT","tier":"open"},{"id":"gpt-4.1","label":"GPT-4.1","description":"Coding, function calling, and structured outputs.","task_type":"decoder","context_window":1047576,"input_price_per_million":2.0,"output_price_per_million":8.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-4.1-mini","label":"GPT-4.1 mini","description":"Fast coding, function calling, and JSON outputs.","task_type":"decoder","context_window":1047576,"input_price_per_million":0.4,"output_price_per_million":1.6,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-4.1-nano","label":"GPT-4.1 nano","description":"Lowest-latency GPT; high-throughput classification and extraction.","task_type":"decoder","context_window":1047576,"input_price_per_million":0.1,"output_price_per_million":0.4,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-4o","label":"GPT-4o","description":"Multimodal reasoning, vision, and tool use.","task_type":"decoder","context_window":128000,"input_price_per_million":2.5,"output_price_per_million":10.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-4o-mini","label":"GPT-4o mini","description":"Fast multimodal reasoning, vision, and structured outputs.","task_type":"decoder","context_window":128000,"input_price_per_million":0.15,"output_price_per_million":0.6,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5-mini","label":"GPT-5 mini","description":"Fast general reasoning, coding, and function calling.","task_type":"decoder","context_window":400000,"input_price_per_million":0.25,"output_price_per_million":2.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5-nano","label":"GPT-5 nano","description":"Lowest-latency GPT-5; high-throughput classification tasks.","task_type":"decoder","context_window":400000,"input_price_per_million":0.05,"output_price_per_million":0.4,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5.1","label":"GPT-5.1","description":"Advanced reasoning, coding, and long-context tasks.","task_type":"decoder","context_window":400000,"input_price_per_million":1.25,"output_price_per_million":10.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5.4","label":"GPT-5.4","description":"Complex reasoning, coding, and agentic tool use.","task_type":"decoder","context_window":1047576,"input_price_per_million":2.5,"output_price_per_million":15.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5.4-mini","label":"GPT-5.4 mini","description":"Fast coding, tool use, and structured outputs.","task_type":"decoder","context_window":400000,"input_price_per_million":0.75,"output_price_per_million":4.5,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5.4-nano","label":"GPT-5.4 nano","description":"Ultra-fast, low-cost classification and structured extraction.","task_type":"decoder","context_window":1047576,"input_price_per_million":0.2,"output_price_per_million":1.25,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"gpt-5.5","label":"GPT-5.5","description":"Frontier reasoning and multi-step agentic workflows.","task_type":"decoder","context_window":400000,"input_price_per_million":5.0,"output_price_per_million":30.0,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"openai/gpt-oss-120b","label":"GPT-OSS 120B","description":"Complex instruction-following, coding, structured outputs.","task_type":"decoder","context_window":131072,"input_price_per_million":0.15,"output_price_per_million":0.6,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Apache-2.0","tier":"open"},{"id":"openai/gpt-oss-20b","label":"GPT-OSS 20B","description":"Fast instruction-following and everyday reasoning tasks.","task_type":"decoder","context_window":131072,"input_price_per_million":0.07,"output_price_per_million":0.3,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Apache-2.0","tier":"open"},{"id":"text-embedding-3-large","label":"text-embedding-3-large","description":"Openai · 8,191 ctx","task_type":"encoder","context_window":8191,"input_price_per_million":0.13,"output_price_per_million":null,"supports_inference":false,"is_chat_model":false,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"text-embedding-3-small","label":"text-embedding-3-small","description":"Openai · 8,191 ctx","task_type":"encoder","context_window":8191,"input_price_per_million":0.02,"output_price_per_million":null,"supports_inference":false,"is_chat_model":false,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"text-embedding-ada-002","label":"text-embedding-ada-002","description":"Openai · 8,191 ctx","task_type":"encoder","context_window":8191,"input_price_per_million":0.1,"output_price_per_million":null,"supports_inference":false,"is_chat_model":false,"supports_on_demand_inference":false,"supports_training":false,"license":"Proprietary","tier":"enterprise"},{"id":"Qwen/Qwen2.5-Coder-0.5B","label":"Qwen2.5 Coder 0.5B","description":"Code completion; very low-latency deployment.","task_type":"decoder","context_window":32768,"input_price_per_million":0.1,"output_price_per_million":0.1,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-1.7B-Base","label":"Qwen3 1.7B Base","description":"Qwen3 1.7B pretrained base — PostTrainBench post-training target.","task_type":"decoder","context_window":32768,"input_price_per_million":0.1,"output_price_per_million":0.1,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-235B-A22B-Instruct-2507","label":"Qwen3 235B A22B Instruct","description":"Complex reasoning, coding, extended thinking support.","task_type":"decoder","context_window":262144,"input_price_per_million":1.2,"output_price_per_million":1.2,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":false,"supports_training":false,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-32B","label":"Qwen3 32B","description":"Coding, math, and reasoning with thinking-mode support.","task_type":"decoder","context_window":40960,"input_price_per_million":0.9,"output_price_per_million":0.9,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-4B-Base","label":"Qwen3 4B Base","description":"Qwen3 4B pretrained base — PostTrainBench post-training target.","task_type":"decoder","context_window":32768,"input_price_per_million":0.15,"output_price_per_million":0.15,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-4B-Instruct-2507","label":"Qwen3 4B Instruct","description":"Fast instruction-following; high-throughput use cases.","task_type":"decoder","context_window":262144,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":false,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3-8B","label":"Qwen3 8B","description":"Coding, math, structured extraction; strong multilingual.","task_type":"decoder","context_window":40960,"input_price_per_million":0.2,"output_price_per_million":0.2,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3.5-9B","label":"Qwen3.5 9B","description":"Balanced reasoning, coding, and multilingual generation.","task_type":"decoder","context_window":32768,"input_price_per_million":0.3,"output_price_per_million":0.3,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"Qwen/Qwen3.6-27B","label":"Qwen3.6 27B","description":"Strong reasoning, coding, and multilingual generation.","task_type":"decoder","context_window":32768,"input_price_per_million":0.6,"output_price_per_million":0.6,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"},{"id":"HuggingFaceTB/SmolLM3-3B-Base","label":"SmolLM3 3B Base","description":"SmolLM3 3B pretrained base — PostTrainBench post-training target.","task_type":"decoder","context_window":32768,"input_price_per_million":0.15,"output_price_per_million":0.15,"supports_inference":true,"is_chat_model":true,"supports_on_demand_inference":true,"supports_training":true,"license":"Apache-2.0","tier":"open"}]}