LLM pricing dataset
AI model pricing table
Browse 281 crawler-readable model rows across 38 providers. Prices are USD per 1,000,000 tokens.
Models281
Providers38
Last updated2026-06-21
| Model | Provider | Input $/M | Output $/M | Context | Tier | Best for |
|---|---|---|---|---|---|---|
| Hunyuan-Lite | Tencent | $0 | $0 | 256,000 | Budget | Free Tencent Cloud tier for high-volume chat and classification |
| Owl Alpha | OpenRouter | $0 | $0 | 1,048,756 | Flagship | Free alpha — high-performance agentic foundation model with native tool use and ~1.05M context. Prompts/completions may be logged. |
| Gemma 2 9B | $0.030 | $0.060 | 8,192 | Open Source | Smaller open-weight Gemma 2 for edge and on-device | |
| Nova Micro | Amazon | $0.035 | $0.140 | 128,000 | Budget | Cheapest text-only Bedrock model for high-volume workloads |
| Command R7B | Cohere | $0.037 | $0.150 | 128,000 | Budget | Ultra-cheap, low-latency high-volume tasks |
| Ministral 3B | Mistral | $0.040 | $0.040 | 128,000 | Budget | Ultra-cheap 3B model for classification and routing |
| Phi-4 Mini | Microsoft | $0.040 | $0.100 | 128,000 | Budget | Compact 3.8B Phi-4 sibling for low-latency reasoning |
| Llama 3 8B | Meta | $0.050 | $0.080 | 8,192 | Open Source | Previous-gen 8B open-weight for chat and edge |
| Llama 3.1 8B | Meta | $0.050 | $0.080 | 128,000 | Open Source | Open-weight 8B for low-latency chat and edge |
| Llama 3.1 8B Instant (Groq) | Groq | $0.050 | $0.080 | 128,000 | Budget | Ultra-fast LPU-hosted 8B for sub-second responses |
| Gemma 3 12B | $0.050 | $0.100 | 128,000 | Open Source | Mid-size open-weight Gemma 3 for self-hosted chat and reasoning | |
| Qwen3 Turbo | Qwen | $0.050 | $0.200 | 131,072 | Budget | High-throughput cost-efficient inference |
| GLM-4.5 Flash | Zhipu | $0.050 | $0.200 | 128,000 | Budget | High-volume classification and retrieval |
| GPT-5.4 Nano | OpenAI | $0.050 | $0.400 | 128,000 | Budget | Classification, routing, simple tasks |
| Qwen3 Flash | Qwen | $0.050 | $0.400 | 128,000 | Budget | Ultra-cheap classification, chat, and high-volume tasks |
| GPT-5 Nano | OpenAI | $0.050 | $0.400 | 400,000 | Budget | Fast, low-cost OpenAI model for simple or high-volume tasks |
| Nova Canvas | Amazon | $0.060 | $0.060 | 0 | Mid-tier | Image generation and editing on Amazon Bedrock |
| Hunyuan-Standard | Tencent | $0.060 | $0.130 | 32,000 | Budget | Default Tencent Cloud model for everyday chat and RAG |
| Hy3 Lite | Tencent | $0.060 | $0.180 | 128,000 | Budget | Ultra-fast MoE for chat, classification, and high-volume apps |
| MiMo Lite | Xiaomi | $0.060 | $0.180 | 128,000 | Budget | Ultra-affordable entry model for high-volume apps |
| Nova Lite | Amazon | $0.060 | $0.240 | 300,000 | Budget | Affordable multimodal Bedrock model — text, vision, video understanding |
| Phi-4 | Microsoft | $0.065 | $0.140 | 16,000 | Budget | Low-latency coding, reasoning, and chat on Azure AI Foundry |
| Grok Image | xAI | $0.070 | $0.070 | 0 | Open Source | Aurora image generation — photorealistic text-to-image |
| abab6.5t-chat | MiniMax | $0.070 | $0.070 | 8,192 | Budget | Tiny abab6.5 variant for ultra-low-latency apps |
| Mistral 7B Instruct | Mistral | $0.070 | $0.070 | 32,000 | Open Source | Classic open-weight 7B instruct model |
| Gemma 2 27B | $0.070 | $0.140 | 8,192 | Open Source | Open-weight 27B model for self-hosted general purpose | |
| Gemini 2.0 Flash-Lite | $0.075 | $0.300 | 1,000,000 | Budget | Ultra-affordable high-volume tasks | |
| Gemini 1.5 Flash | $0.075 | $0.300 | 1,000,000 | Budget | Fast, affordable previous-gen model with 1M context | |
| Nemotron Mini 4B Instruct | Nvidia | $0.080 | $0.080 | 4,096 | Budget | Tiny on-device Nemotron for edge inference |
| Gemma 3 27B | $0.080 | $0.160 | 128,000 | Open Source | Open-weight multimodal model for text and vision tasks | |
| MiniMax Text-Lite | MiniMax | $0.080 | $0.240 | 128,000 | Budget | Ultra-affordable entry model for chat and high-volume apps |
| Llama 4 Scout (Fireworks AI) | Fireworks AI | $0.090 | $0.250 | 1,000,000 | Budget | Fireworks-hosted Llama 4 Scout — ultra-fast inference |
| Ministral 8B | Mistral | $0.100 | $0.100 | 128,000 | Budget | Edge-class 8B model — low-latency on-device chat |
| Mistral Embed 2 | Mistral | $0.100 | $0.100 | 32,000 | Budget | Multilingual embeddings for RAG and clustering |
| Mistral Moderation | Mistral | $0.100 | $0.100 | 32,000 | Budget | Content moderation across 9 safety categories |
| Granite 3.3 2B Instruct | IBM | $0.100 | $0.100 | 128,000 | Budget | Ultra-cheap on-device/edge inference, classification, summarization on watsonx.ai |
| Embed v3 English | Cohere | $0.100 | $0.100 | 512 | Budget | 1024-dim English embeddings for semantic search and RAG |
| Granite Vision 3.2 2B | IBM | $0.100 | $0.100 | 16,000 | Budget | Compact vision-language model for document and chart understanding |
| Yi-Coder 9B | 01.AI | $0.100 | $0.100 | 128,000 | Open Source | Open-weight 9B code model — competitive with 33B on benchmarks |
| Llama 3.1 8B (Cerebras) | Cerebras | $0.100 | $0.100 | 128,000 | Budget | Wafer-scale 8B for ultra-low-latency chat |
| Llama 4 Scout | Meta | $0.100 | $0.250 | 10,000,000 | Open Source | Free self-host, efficient |
| Grok 4.1 Mini | xAI | $0.100 | $0.300 | 1,000,000 | Budget | Smallest 4.1 variant — RAG and high-volume classification |
| Devstral Small | Mistral | $0.100 | $0.300 | 128,000 | Budget | Open-weight agentic coder for IDEs and local dev |
| Nemotron Nano 9B v2 | Nvidia | $0.100 | $0.300 | 128,000 | Budget | Cheapest fast Nemotron — high-throughput agents and RAG |
| MiMo V2.5 Lite | Xiaomi | $0.100 | $0.300 | 128,000 | Budget | Fast, low-cost variant for chat, classification, and routing |
| Llama 4 Scout (Together AI) | Together AI | $0.100 | $0.300 | 1,000,000 | Budget | Hosted open-weight inference — fast, affordable general purpose |
| Phi-5 | Microsoft | $0.100 | $0.300 | 256,000 | Mid-tier | Next-gen Phi small-model flagship — strong quality per dollar |
| GPT-4.1 Nano | OpenAI | $0.100 | $0.400 | 1,000,000 | Budget | High-volume, low-latency workloads |
| Gemini 2.5 Flash-Lite | $0.100 | $0.400 | 1,000,000 | Budget | Cheapest major provider option | |
| Gemini 2.0 Flash | $0.100 | $0.400 | 1,000,000 | Budget | Previous-gen fast model | |
| Hunyuan-TurboS | Tencent | $0.110 | $0.420 | 256,000 | Mid-tier | Hybrid Transformer-Mamba balance of speed and quality |
| Nano (STT) | AssemblyAI | $0.120 | $0 | 0 | Budget | Budget STT: $0.12 per audio hour |
| Embed v4 Multilingual | Cohere | $0.120 | $0.120 | 4,096 | Budget | 1024-dim multilingual embeddings across 100+ languages |
| Mistral Small 3.1 (Fireworks AI) | Fireworks AI | $0.130 | $0.500 | 128,000 | Budget | Fireworks-hosted Mistral Small 3.1 for high-volume agents |
| Yi-Lightning | 01.AI | $0.140 | $0.140 | 16,384 | Flagship | Low-latency flagship for chat and reasoning |
| abab6.5s-chat | MiniMax | $0.140 | $0.140 | 245,000 | Budget | Lighter, faster abab6.5 sibling for high-volume chat |
| DeepSeek V2.5 | DeepSeek | $0.140 | $0.280 | 128,000 | Mid-tier | Previous-gen DeepSeek for general chat and coding |
| DeepSeek Coder V2 | DeepSeek | $0.140 | $0.280 | 128,000 | Mid-tier | Code-specialized MoE for IDE completion and repo-scale tasks |
| DeepSeek V4 Flash | DeepSeek | $0.140 | $0.550 | 128,000 | Budget | 284B MoE (13B active), high-throughput chat/RAG/extraction |
| Hunyuan-T1 | Tencent | $0.140 | $0.550 | 128,000 | Reasoning | Hybrid Mamba-Transformer reasoning model with strong math and coding |
| Codestral Embed | Mistral | $0.150 | $0.150 | 8,192 | Budget | Code embeddings for retrieval and semantic search |
| Pixtral 12B | Mistral | $0.150 | $0.150 | 128,000 | Open Source | Open-weight 12B vision-language model |
| Mistral NeMo | Mistral | $0.150 | $0.150 | 128,000 | Open Source | Open-weight 12B model co-developed with NVIDIA |
| Mistral Small 3.1 | Mistral | $0.150 | $0.600 | 128,000 | Budget | High-throughput budget tasks |
| Command R | Cohere | $0.150 | $0.600 | 128,000 | Mid-tier | Production RAG at budget pricing |
| Kimi K2 Mini | Moonshot | $0.150 | $0.600 | 128,000 | Budget | Distilled K2 for low-latency, high-throughput chat |
| GPT-4o Mini (GitHub Models) | GitHub Models | $0.150 | $0.600 | 128,000 | Budget | Affordable Azure-hosted small multimodal model |
| QwQ 32B | Qwen | $0.150 | $0.600 | 131,072 | Reasoning | Open-weight reasoning model with extended chain-of-thought |
| Phi-3.5 MoE | Microsoft | $0.160 | $0.640 | 131,072 | Mid-tier | 16x3.8B sparse MoE — strong quality-per-dollar on Azure AI Foundry |
| Qwen 2.5 Coder 32B | Qwen | $0.180 | $0.180 | 131,072 | Open Source | Code-specialized 32B for completion and refactoring |
| Llama 4 Maverick (Fireworks AI) | Fireworks AI | $0.180 | $0.500 | 1,000,000 | Mid-tier | Fireworks-hosted Llama 4 Maverick — low-latency reasoning |
| Qwen 2.5 32B Instruct | Qwen | $0.200 | $0.200 | 131,072 | Open Source | Mid-size Qwen 2.5 for production agents |
| Sonar Small Online | Perplexity | $0.200 | $0.200 | 127,000 | Budget | Cheapest grounded search model with citations |
| Granite Code 8B Instruct | IBM | $0.200 | $0.200 | 128,000 | Budget | Code-specialized 8B for enterprise IDE integrations on watsonx.ai |
| Yi-2 Lightning | 01.AI | $0.200 | $0.200 | 32,768 | Flagship | Next-gen Yi — low-latency bilingual flagship for chat and reasoning |
| Granite 4 8B Instruct | IBM | $0.200 | $0.200 | 256,000 | Open Source | Hybrid Mamba-Transformer enterprise instruct on watsonx.ai |
| Jamba 1.7 Mini | AI21 | $0.200 | $0.400 | 256,000 | Budget | Cost-efficient hybrid SSM for long-context RAG |
| Grok 4.1 Fast | xAI | $0.200 | $0.500 | 2,000,000 | Budget | Ultra-cheap with massive 2M context |
| Llama 4 Maverick | Meta | $0.200 | $0.600 | 1,000,000 | Open Source | Free self-host, strong general purpose |
| Mistral Saba | Mistral | $0.200 | $0.600 | 32,000 | Mid-tier | Arabic + South-Asian language specialist |
| Llama 4 Maverick (Together AI) | Together AI | $0.200 | $0.600 | 1,000,000 | Mid-tier | Hosted open-weight inference — strong reasoning and coding |
| Llama 3.3 70B | Meta | $0.200 | $0.600 | 128,000 | Open Source | 70B instruction-tuned open-weight — strong general purpose |
| Llama 3.1 70B | Meta | $0.200 | $0.600 | 128,000 | Open Source | Open-weight 70B for production agents and reasoning |
| DeepSeek Coder V3 | DeepSeek | $0.200 | $0.800 | 256,000 | Flagship | Code-specialized MoE for repo-scale completion and refactoring |
| GLM-4.5 Air | Zhipu | $0.200 | $1.10 | 128,000 | Mid-tier | Lightweight MoE for cost-sensitive agentic workloads |
| Grok Code Fast 1 | xAI | $0.200 | $1.50 | 256,000 | Mid-tier | Coding-specialized Grok for IDEs and agentic dev |
| Hunyuan-Standard-256K | Tencent | $0.210 | $0.840 | 256,000 | Mid-tier | Long-context variant of Hunyuan Standard for document QA |
| Llama 3.3 70B (DeepInfra) | DeepInfra | $0.230 | $0.400 | 128,000 | Mid-tier | DeepInfra-hosted Llama 3.3 70B at aggressive per-token pricing |
| Mixtral 8x7B | Mistral | $0.240 | $0.240 | 32,000 | Open Source | Sparse MoE 8x7B — efficient open-weight workhorse |
| Mixtral 8x7B (Groq) | Groq | $0.240 | $0.240 | 32,000 | Mid-tier | LPU-hosted Mixtral 8x7B with extreme throughput |
| Qwen3 32B (Fireworks AI) | Fireworks AI | $0.250 | $0.700 | 128,000 | Mid-tier | Fireworks-hosted Qwen3 32B optimized for throughput |
| Llama 4.1 70B | Meta | $0.250 | $0.700 | 1,000,000 | Open Source | Refreshed 70B with 1M context — strong general purpose |
| MiMo Text | Xiaomi | $0.250 | $0.750 | 256,000 | Mid-tier | Mid-tier reasoning, coding, and long-context RAG |
| Claude 3 Haiku | Anthropic | $0.250 | $1.25 | 200,000 | Budget | Cheapest Claude 3 — high-volume chat and extraction |
| Gemini 3.1 Flash-Lite | $0.250 | $1.50 | 1,000,000 | Budget | Budget with newer capabilities | |
| GPT-5 Mini | OpenAI | $0.250 | $2.00 | 400,000 | Mid-tier | Lower-cost GPT-5 for strong general performance |
| DeepSeek V3 | DeepSeek | $0.270 | $1.10 | 128,000 | Flagship | 671B MoE flagship — strong coding and reasoning at open-weight prices |
| DeepSeek V3.2 | DeepSeek | $0.280 | $0.420 | 128,000 | Budget | Best value overall, 90% cache discount |
| Qwen3 32B (Groq) | Groq | $0.290 | $0.590 | 128,000 | Mid-tier | LPU-accelerated Qwen3 32B — extreme throughput for agents |
| Speechmatics Standard | Speechmatics | $0.300 | $0 | 0 | Budget | STT: $0.30 per audio hour — 50+ languages with diarization |
| abab6.5-chat | MiniMax | $0.300 | $0.300 | 245,000 | Mid-tier | Trillion-parameter MoE chat model with long context |
| Llama 3 70B | Meta | $0.300 | $0.400 | 8,192 | Open Source | Previous-gen Llama 3 70B instruct |
| Grok 3 Mini | xAI | $0.300 | $0.500 | 131,072 | Budget | Budget Grok 3 for chat and lightweight agents |
| Nemotron 3 Super 120B | Nvidia | $0.300 | $0.800 | 262,000 | Reasoning | Hybrid Mamba-Transformer MoE — agentic reasoning, long context |
| Qwen3 32B (Together AI) | Together AI | $0.300 | $0.800 | 128,000 | Mid-tier | Hosted Qwen3 32B for production agents and tool use |
| Grok 4 Mini | xAI | $0.300 | $0.900 | 256,000 | Budget | Cheaper Grok 4 distilled for high-throughput chat and tools |
| Codestral 25.10 | Mistral | $0.300 | $0.900 | 256,000 | Mid-tier | Code generation, completion, and FIM across 80+ languages |
| Sarvam-1 | Sarvam AI | $0.300 | $0.900 | 8,192 | Budget | Compact Indic foundation model hosted in India |
| Hunyuan Pro | Tencent | $0.300 | $1.00 | 32,000 | Flagship | Production Hunyuan flagship for everyday agents and RAG |
| MiniMax M2 | MiniMax | $0.300 | $1.20 | 1,000,000 | Flagship | 230B MoE optimized for agentic coding and end-to-end tool use |
| Gemini 2.5 Flash | $0.300 | $2.50 | 1,000,000 | Budget | High-volume, budget apps | |
| Qwen 2.5 72B Instruct | Qwen | $0.350 | $0.400 | 131,072 | Open Source | Open-weight 72B instruct — strong multilingual reasoning |
| Codestral 26 | Mistral | $0.350 | $1.00 | 256,000 | Mid-tier | Refreshed Codestral for completion, FIM, and 80+ languages |
| MiniMax Text-01 | MiniMax | $0.350 | $1.05 | 256,000 | Flagship | Long-context text flagship — strong reasoning and RAG |
| Hy3 Preview | Tencent | $0.350 | $1.05 | 128,000 | Flagship | Frontier preview — strong reasoning, coding, and agentic workflows |
| MiMo Audio | Xiaomi | $0.350 | $1.05 | 32,000 | Flagship | End-to-end speech + text model for voice agents and audio reasoning |
| Universal-2 | AssemblyAI | $0.370 | $0 | 0 | Flagship | Speech-to-text: $0.37 per audio hour |
| Llama 3.3 Nemotron Super 49B | Nvidia | $0.400 | $0.400 | 128,000 | Flagship | NVIDIA-tuned Llama for tool use and instruction following |
| Llama 3.1 Nemotron 70B Instruct | Nvidia | $0.400 | $0.400 | 128,000 | Flagship | NVIDIA-tuned Llama 70B with strong arena scores |
| abab7-chat | MiniMax | $0.400 | $0.400 | 256,000 | Mid-tier | Next-gen abab chat model with extended context and stronger reasoning |
| Qwen3 32B (Cerebras) | Cerebras | $0.400 | $0.800 | 128,000 | Mid-tier | Wafer-scale Qwen3 32B at 2000+ tokens/sec inference |
| Reka Edge | Reka | $0.400 | $1.00 | 64,000 | Budget | On-device / edge-deployable multimodal model |
| Qwen3 Plus | Qwen | $0.400 | $1.20 | 256,000 | Mid-tier | Balanced quality and price for production apps |
| Hunyuan-Turbo | Tencent | $0.400 | $1.20 | 256,000 | Flagship | Production MoE flagship — fast, multilingual, function calling |
| MiMo Vision | Xiaomi | $0.400 | $1.20 | 128,000 | Flagship | Vision-language MoE for multimodal reasoning and image understanding |
| Grok 5 Mini | xAI | $0.400 | $1.50 | 1,000,000 | Mid-tier | Cheaper distilled Grok 5 for high-throughput chat and tools |
| GPT-5.4 Mini | OpenAI | $0.400 | $1.60 | 128,000 | Mid-tier | Production apps, moderate complexity |
| GPT-4.1 Mini | OpenAI | $0.400 | $1.60 | 1,000,000 | Mid-tier | Structured tasks, cost-efficient coding |
| Mistral Medium 3 | Mistral | $0.400 | $2.00 | 131,000 | Mid-tier | Balanced workloads, function calling |
| Devstral Medium | Mistral | $0.400 | $2.00 | 256,000 | Mid-tier | Agentic coding — long-context repo edits and tool use |
| Kimi K1.5 | Moonshot | $0.400 | $2.00 | 128,000 | Mid-tier | Long-context multimodal reasoning with RL post-training |
| MiniMax M1 | MiniMax | $0.400 | $2.20 | 1,000,000 | Reasoning | 456B MoE reasoning model with 1M-token context — first open M-series |
| MiniMax VL-01 | MiniMax | $0.450 | $1.35 | 128,000 | Flagship | Vision-language MoE for multimodal reasoning and image understanding |
| MiMo V2.5 | Xiaomi | $0.450 | $1.35 | 256,000 | Flagship | Open-weight MoE flagship — strong reasoning, coding, and RAG |
| Granite 3.3 8B Instruct | IBM | $0.500 | $0.500 | 128,000 | Open Source | Enterprise RAG, summarization, function calling on watsonx.ai |
| Yi-Medium | 01.AI | $0.500 | $0.500 | 16,384 | Mid-tier | Balanced cost-quality model |
| Yi-VL Plus | 01.AI | $0.500 | $0.500 | 16,384 | Mid-tier | Bilingual vision-language model for charts and OCR |
| Jamba Instruct | AI21 | $0.500 | $0.700 | 256,000 | Mid-tier | Instruction-tuned variant for enterprise deployments |
| Magistral Small | Mistral | $0.500 | $1.50 | 128,000 | Mid-tier | Open-weight 24B reasoning model — math, science, code |
| Sarvam-M | Sarvam AI | $0.500 | $1.50 | 32,768 | Mid-tier | Indic-first 24B reasoning model across 10+ Indian languages |
| Hunyuan-Large | Tencent | $0.500 | $1.50 | 256,000 | Open Source | 389B open-weight MoE — 52B active params, strong general reasoning |
| Hunyuan-Vision | Tencent | $0.500 | $1.50 | 32,000 | Flagship | Multimodal vision-language model for image + text reasoning |
| MiMo Coder | Xiaomi | $0.500 | $1.50 | 128,000 | Flagship | Code-specialized MiMo for repo-scale coding and refactoring |
| Aya Expanse 32B | Cohere | $0.500 | $1.50 | 128,000 | Mid-tier | 23-language open-weight model from Cohere For AI |
| Qwen3 Vision | Qwen | $0.500 | $1.50 | 256,000 | Flagship | Vision-language Qwen3 for charts, OCR, and UI understanding |
| Mistral Large 3 | Mistral | $0.500 | $2.00 | 131,000 | Flagship | European provider, multilingual |
| GLM-Z1 | Zhipu | $0.500 | $2.00 | 128,000 | Reasoning | Reasoning-tuned GLM with extended chain-of-thought |
| MiniMax M3 | MiniMax | $0.500 | $2.20 | 4,000,000 | Flagship | Flagship MoE with 4M-token context — agentic reasoning and tool use |
| Kimi Latest | Moonshot | $0.500 | $2.50 | 200,000 | Mid-tier | Always-updated production endpoint tracking the newest Kimi |
| Gemini 3 Flash | $0.500 | $3.00 | 1,000,000 | Mid-tier | Fast reasoning, good value | |
| Hunyuan-Code | Tencent | $0.550 | $1.65 | 128,000 | Flagship | Code-specialized Hunyuan for repo-scale coding and refactoring |
| Hunyuan-Role | Tencent | $0.550 | $1.65 | 32,000 | Mid-tier | Role-play tuned variant for character agents and companions |
| Hunyuan-FunctionCall | Tencent | $0.550 | $1.65 | 32,000 | Mid-tier | Tool-use tuned variant optimized for function calling pipelines |
| DeepSeek R1 | DeepSeek | $0.550 | $2.19 | 128,000 | Reasoning | Reasoning at budget prices |
| DeepSeek V4 Pro | DeepSeek | $0.550 | $2.20 | 256,000 | Flagship | 1.6T MoE flagship (49B active), frontier reasoning at open-weight prices |
| Kimi K2 Instruct | Moonshot | $0.550 | $2.20 | 262,144 | Flagship | Open-weight 1T MoE instruct model — coding, tool use, agents |
| MiMo Reasoner | Xiaomi | $0.550 | $2.20 | 128,000 | Reasoning | Reasoning-tuned MoE for math, science, and multi-step problems |
| Llama 3.3 70B (Groq) | Groq | $0.590 | $0.790 | 128,000 | Mid-tier | LPU-accelerated Llama 3.3 70B — among the fastest tokens/sec available |
| Llama 3.3 70B (SambaNova) | SambaNova | $0.600 | $1.20 | 128,000 | Mid-tier | RDU-accelerated Llama 3.3 70B for fast enterprise inference |
| Pharia-1-LLM-7B-Control | Aleph Alpha | $0.600 | $1.80 | 8,192 | Budget | EU-sovereign multilingual instruct model — GDPR + BSI C5 compliant |
| GLM-4.6 | Zhipu | $0.600 | $2.20 | 200,000 | Flagship | Zhipu flagship reasoning and coding model |
| Mistral Large 4 | Mistral | $0.600 | $2.40 | 256,000 | Flagship | Next-gen European flagship — multilingual and function calling |
| MiniMax M4 | MiniMax | $0.600 | $2.40 | 4,000,000 | Premium | Successor to M3 — 4M context with stronger agentic reasoning |
| Kimi Coder | Moonshot | $0.650 | $2.60 | 262,144 | Flagship | Repo-scale code generation, refactoring, and agentic dev |
| GLM-4-Plus | Zhipu | $0.700 | $2.10 | 128,000 | Flagship | Premium GLM variant for complex reasoning and tool use |
| MiMo V3 | Xiaomi | $0.700 | $2.10 | 512,000 | Flagship | Next-gen open-weight MoE — long-context reasoning and tool use |
| Kimi Math | Moonshot | $0.700 | $2.80 | 128,000 | Mid-tier | Specialized math and theorem-proving variant |
| DeepSeek R2 | DeepSeek | $0.700 | $2.80 | 256,000 | Reasoning | Successor to R1 — frontier open-weight reasoning at budget prices |
| Qwen 2.5 Coder 32B (Together AI) | Together AI | $0.800 | $0.800 | 131,072 | Mid-tier | Hosted Qwen 2.5 Coder 32B for IDE-grade completions |
| Reka Flash | Reka | $0.800 | $2.00 | 128,000 | Mid-tier | Fast multimodal model for production agents |
| MiMo Pro | Xiaomi | $0.800 | $2.40 | 256,000 | Flagship | Production MoE flagship with extended context and tool use |
| Llama 4 Behemoth | Meta | $0.800 | $2.40 | 256,000 | Premium | 2T-parameter teacher model — frontier open-weight quality |
| Nemotron 4 Ultra | Nvidia | $0.800 | $2.40 | 256,000 | Flagship | Frontier Nemotron MoE for agentic reasoning and tool use |
| GLM-5 | Zhipu | $0.800 | $3.00 | 256,000 | Flagship | Next-gen GLM flagship for reasoning, coding, and tool use |
| Nova Pro | Amazon | $0.800 | $3.20 | 300,000 | Flagship | High-quality text, vision, and tool use via Amazon Bedrock |
| Kimi K1.5 Long | Moonshot | $0.800 | $3.20 | 2,000,000 | Flagship | 2M-token context for full codebases and document corpora |
| Kimi Audio | Moonshot | $0.800 | $3.20 | 64,000 | Mid-tier | Speech-in / speech-out multimodal audio model |
| o5-mini | OpenAI | $0.800 | $3.20 | 256,000 | Reasoning | Compact o5 reasoning for everyday problem solving |
| Claude 3.5 Haiku | Anthropic | $0.800 | $4.00 | 200,000 | Budget | Fast, affordable Claude 3.5 for chat and classification |
| Llama 3.3 70B (Cerebras) | Cerebras | $0.850 | $1.20 | 128,000 | Mid-tier | Wafer-scale Llama 3.3 70B — 2000+ tokens/sec inference |
| Llama 3.3 70B (Together AI) | Together AI | $0.880 | $0.880 | 128,000 | Mid-tier | Hosted Llama 3.3 70B with fast turbo endpoints |
| DeepSeek V3 (Together AI) | Together AI | $0.900 | $0.900 | 128,000 | Budget | Hosted DeepSeek V3 with fast autoscaling inference |
| Llama 3.3 70B (Fireworks AI) | Fireworks AI | $0.900 | $0.900 | 128,000 | Mid-tier | Fireworks-hosted Llama 3.3 70B — low-latency inference |
| Qwen 2.5 Coder 32B (Fireworks AI) | Fireworks AI | $0.900 | $0.900 | 131,072 | Mid-tier | Fireworks-hosted Qwen 2.5 Coder 32B for agentic dev tools |
| MiniMax Pro | MiniMax | $0.900 | $2.70 | 256,000 | Flagship | Max-capability variant with extended-context reasoning and tool use |
| Hy3 Pro | Tencent | $0.900 | $2.70 | 256,000 | Flagship | Production-grade MoE flagship for complex agents and long-context RAG |
| Kimi Vision | Moonshot | $0.900 | $3.60 | 128,000 | Flagship | Multimodal vision-language model for charts, OCR, and UI |
| Kimi K2.6 | Moonshot | $0.950 | $4.00 | 262,144 | Flagship | Frontier-scale coding, long-context agentic workflows |
| Mistral OCR | Mistral | $1.00 | $1.00 | 0 | Mid-tier | Document OCR with layout and table extraction |
| Sonar | Perplexity | $1.00 | $1.00 | 127,000 | Budget | Fast grounded web search and everyday Q&A with citations |
| Grok Voice | xAI | $1.00 | $4.00 | 32,000 | Mid-tier | Realtime voice-mode model for spoken interactions |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | 200,000 | Budget | Classification, chat, extraction |
| Qwen3 Coder | Qwen | $1.00 | $5.00 | 1,000,000 | Flagship | Best-in-class open-weight coding and tool use |
| Sonar Reasoning | Perplexity | $1.00 | $5.00 | 127,000 | Reasoning | Chain-of-thought reasoning grounded in real-time web search |
| Grok 1.5 | xAI | $1.00 | $5.00 | 128,000 | Mid-tier | Original Grok 1.5 — extended-context chat and reasoning |
| Claude Haiku 4.6 | Anthropic | $1.00 | $5.00 | 200,000 | Budget | Faster Haiku refresh — chat, classification, extraction |
| Speechmatics Enhanced | Speechmatics | $1.04 | $0 | 0 | Flagship | Best-in-class STT: $1.04 per audio hour, on-prem available |
| o4-mini | OpenAI | $1.10 | $4.40 | 128,000 | Reasoning | Strong problem-solving and logical deduction |
| o3-mini | OpenAI | $1.10 | $4.40 | 128,000 | Reasoning | Strong problem-solving and logical deduction |
| MiMo V2.5 Pro | Xiaomi | $1.20 | $3.60 | 256,000 | Flagship | Flagship MiMo with extended context and agentic tool use |
| Mixtral 8x22B (Together AI) | Together AI | $1.20 | $3.60 | 64,000 | Flagship | Hosted sparse MoE — multilingual and long-context |
| Kimi K2 Thinking | Moonshot | $1.20 | $5.00 | 262,144 | Premium | Deep step-by-step reasoning across math, science, and code |
| Kimi K3 | Moonshot | $1.20 | $5.00 | 512,000 | Premium | Next-gen Kimi flagship — agentic coding, 512K context, tool use |
| Gemini 1.5 Pro | $1.25 | $5.00 | 2,000,000 | Flagship | Previous-gen long-context multimodal flagship | |
| Gemini 2.5 Pro | $1.25 | $10.00 | 2,000,000 | Flagship | Coding, analysis, long context | |
| GPT-5 | OpenAI | $1.25 | $10.00 | 400,000 | Flagship | Powerful all-rounder for accuracy, nuance, and multimodal text+image tasks |
| GLM-5.2 | Z.AI | $1.40 | $4.20 | 1,000,000 | Flagship | Z.ai 744B open flagship — #1 open on coding & agent benchmarks, MIT license, 1M context |
| GLM-5.2 | Zhipu | $1.40 | $4.20 | 1,000,000 | Flagship | Z.ai 744B open flagship — #1 open on coding & agent benchmarks, MIT license, 1M context |
| Kimi Researcher | Moonshot | $1.50 | $6.00 | 262,144 | Premium | Autonomous deep-research agent with browsing and citations |
| Qwen3 Max | Qwen | $1.60 | $6.40 | 1,000,000 | Flagship | Frontier reasoning, agentic coding, long context |
| GPT-5.2 | OpenAI | $1.75 | $14.00 | 128,000 | Flagship | Coding, analysis |
| Mixtral 8x22B (Fireworks AI) | Fireworks AI | $1.80 | $5.40 | 64,000 | Flagship | Fireworks-hosted Mixtral 8x22B with speculative decoding |
| Rerank 3 | Cohere | $2.00 | $0 | 4,096 | Mid-tier | Cross-encoder reranking for high-precision RAG ($2 per 1K searches) |
| Reka Core | Reka | $2.00 | $2.00 | 128,000 | Flagship | Multimodal flagship across text, image, audio, and video |
| Magistral Medium 2 | Mistral | $2.00 | $5.00 | 128,000 | Premium | Frontier European reasoning model with extended chain-of-thought |
| Pixtral Large | Mistral | $2.00 | $6.00 | 128,000 | Flagship | Multimodal flagship — vision, charts, document understanding |
| Mixtral 8x22B | Mistral | $2.00 | $6.00 | 64,000 | Open Source | Sparse mixture-of-experts open-weight model |
| Hy3 Max | Tencent | $2.00 | $6.00 | 256,000 | Premium | Maximum-capability Hy3 with extended-context reasoning and tool use |
| Mistral Large 2 | Mistral | $2.00 | $6.00 | 128,000 | Flagship | Previous-gen Mistral flagship — strong reasoning and multilingual |
| GPT-4.1 | OpenAI | $2.00 | $8.00 | 1,000,000 | Flagship | Coding, analysis, long-context tasks |
| o3 | OpenAI | $2.00 | $8.00 | 200,000 | Reasoning | Complex reasoning, research, math |
| Sonar Reasoning Pro | Perplexity | $2.00 | $8.00 | 127,000 | Reasoning | Advanced DeepSeek R1-based reasoning with grounded search |
| Sonar Deep Research | Perplexity | $2.00 | $8.00 | 127,000 | Premium | Expert multi-query research, long-form reports with citations |
| Jamba 1.7 Large | AI21 | $2.00 | $8.00 | 256,000 | Flagship | Hybrid Mamba-Transformer flagship with 256K context |
| GPT-Image 1 Mini | OpenAI | $2.00 | $8.00 | 0 | Mid-tier | Cost-efficient image generation via /v1/images/generations |
| Qwen3.5 Max | Qwen | $2.00 | $8.00 | 1,000,000 | Premium | Next-gen Qwen flagship — frontier reasoning and agentic coding |
| Grok 4.1 | xAI | $2.00 | $10.00 | 2,000,000 | Flagship | Long-context flagship with native tool use and search |
| Grok 2 Vision | xAI | $2.00 | $10.00 | 32,768 | Mid-tier | Multimodal Grok 2 — charts, documents, and screenshots |
| Grok 2 | xAI | $2.00 | $10.00 | 128,000 | Flagship | Previous-gen Grok with strong reasoning and search |
| Gemini 3.1 Pro | $2.00 | $12.00 | 1,000,000 | Premium | Long context, multimodal | |
| Gemini 3 Pro | $2.00 | $12.00 | 2,000,000 | Premium | Frontier multimodal, agentic capabilities | |
| Reka Vision Pro | Reka | $2.50 | $5.00 | 256,000 | Premium | Premium multimodal — video, image, audio, and text reasoning |
| Command A | Cohere | $2.50 | $10.00 | 256,000 | Flagship | Enterprise agents, RAG, and tool use |
| Command R+ | Cohere | $2.50 | $10.00 | 128,000 | Flagship | Advanced RAG and multilingual workflows |
| GPT-4o (GitHub Models) | GitHub Models | $2.50 | $10.00 | 128,000 | Flagship | Azure-hosted GPT-4o via GitHub Models |
| GPT-4o (via Vercel) | Vercel AI Gateway | $2.50 | $10.00 | 128,000 | Flagship | Routed pass-through pricing |
| Nova Premier | Amazon | $2.50 | $12.50 | 1,000,000 | Premium | Largest Bedrock-native model — frontier multimodal reasoning |
| GPT-5.4 | OpenAI | $2.50 | $15.00 | 128,000 | Flagship | General purpose, agents, coding |
| Gemini 3 Pro Vision | $2.50 | $15.00 | 2,000,000 | Premium | Vision-specialized Gemini 3 Pro for charts, documents, and video | |
| Llama 3.1 405B | Meta | $2.70 | $2.70 | 128,000 | Open Source | Largest open-weight Llama — frontier-class quality |
| Yi-Large | 01.AI | $3.00 | $3.00 | 32,768 | Flagship | Bilingual EN/ZH reasoning model |
| DeepSeek R1 (Together AI) | Together AI | $3.00 | $7.00 | 128,000 | Reasoning | Hosted DeepSeek R1 reasoning with autoscaling inference |
| DeepSeek R1 (Fireworks AI) | Fireworks AI | $3.00 | $8.00 | 128,000 | Reasoning | Fireworks-hosted DeepSeek R1 with optimized throughput |
| Hy3 Ultra | Tencent | $3.00 | $9.00 | 512,000 | Premium | Largest Hy3 variant — frontier reasoning and extended context |
| o5 | OpenAI | $3.00 | $12.00 | 256,000 | Reasoning | Next-gen o-series reasoning — math, science, agentic planning |
| Command A Reasoning | Cohere | $3.00 | $12.00 | 256,000 | Reasoning | Reasoning-tuned Command A for enterprise agents and analysis |
| Claude Sonnet 4.6 | Anthropic | $3.00 | $15.00 | 200,000 | Flagship | Production workhorse, instruction following |
| Claude Sonnet 4.5 | Anthropic | $3.00 | $15.00 | 200,000 | Flagship | Recommended default, strong coding |
| Grok 4 | xAI | $3.00 | $15.00 | 256,000 | Flagship | Frontier reasoning and coding |
| Grok 3 | xAI | $3.00 | $15.00 | 131,072 | Flagship | Previous-gen flagship — strong general reasoning |
| Sonar Pro | Perplexity | $3.00 | $15.00 | 200,000 | Flagship | Multi-step research with 2× more citations and deeper sources |
| Claude 3.5 Sonnet (via Vercel) | Vercel AI Gateway | $3.00 | $15.00 | 200,000 | Flagship | Routed via Vercel AI Gateway |
| Claude 3 Sonnet | Anthropic | $3.00 | $15.00 | 200,000 | Mid-tier | Original Claude 3 Sonnet — balanced reasoning and cost |
| Claude 3.5 Sonnet | Anthropic | $3.00 | $15.00 | 200,000 | Flagship | Previous-gen workhorse — coding, vision, and tool use |
| Claude Sonnet 4.7 | Anthropic | $3.00 | $15.00 | 1,000,000 | Flagship | Latest Sonnet — 1M context, stronger coding and tool use |
| Claude Sonnet 3.7 | Anthropic | $3.00 | $15.00 | 200,000 | Flagship | Hybrid reasoning Sonnet 3.7 with extended thinking mode |
| Grok 5 | xAI | $4.00 | $20.00 | 2,000,000 | Premium | Next-gen Grok flagship — agentic reasoning and native search |
| Hume EVI 3 | Hume AI | $4.32 | $0 | 0 | Flagship | Empathic voice (STT+LLM+TTS): $0.072/min ≈ $4.32 per audio hour |
| GPT-4o Realtime | OpenAI | $5.00 | $20.00 | 128,000 | Flagship | Low-latency speech-to-speech Realtime API model |
| Sonar Ultra | Perplexity | $5.00 | $20.00 | 200,000 | Premium | Largest Sonar variant — multi-step research with extensive citations |
| Claude Opus 4.8 | Anthropic | $5.00 | $25.00 | 1,000,000 | Premium | Previous Anthropic flagship — frontier reasoning, agentic coding, 1M context |
| Claude Opus 4.7 | Anthropic | $5.00 | $25.00 | 1,000,000 | Premium | Frontier reasoning, agentic coding, long-context analysis |
| Claude Opus 4.6 | Anthropic | $5.00 | $25.00 | 1,000,000 | Premium | Deep reasoning, long documents, coding |
| Claude Opus 4.5 | Anthropic | $5.00 | $25.00 | 200,000 | Premium | Previous-gen flagship, stable versioned |
| Grok 3 Reasoning | xAI | $5.00 | $25.00 | 131,072 | Premium | Extended chain-of-thought reasoning variant of Grok 3 |
| Gemini 3.1 Ultra | $5.00 | $25.00 | 2,000,000 | Premium | Largest-scale Gemini 3.1 for frontier reasoning and agentic workflows | |
| GPT-5.5 | OpenAI | $5.00 | $30.00 | 256,000 | Flagship | Most intuitive flagship — complex multi-step tasks, agents |
| Grok 4 Heavy | xAI | $6.00 | $30.00 | 256,000 | Premium | Multi-agent parallel reasoning — top scores on ARC-AGI and HLE |
| GPT-5.6 | OpenAI | $6.00 | $36.00 | 400,000 | Flagship | Next-gen flagship — deeper reasoning, multimodal, long-context agents |
| GPT-Image 2 | OpenAI | $10.00 | $40.00 | 0 | Premium | State-of-the-art image generation via /v1/images/generations |
| Claude Fable 5 | Anthropic | $10.00 | $50.00 | 1,000,000 | Premium | Mythos-class flagship — SOTA coding, vision, long-horizon agentic work |
| Octave TTS | Hume AI | $15.00 | $0 | 5,000 | Mid-tier | Emotionally expressive TTS (~$0.015/1K chars) |
| Claude 3 Opus | Anthropic | $15.00 | $75.00 | 200,000 | Premium | Original Opus flagship — complex reasoning and writing |
| GPT-5.5 Pro | OpenAI | $30.00 | $180.00 | 256,000 | Premium | Frontier reasoning, research, agentic workflows |
| GPT-5.4 Pro | OpenAI | $30.00 | $180.00 | 256,000 | Premium | Highly complex tasks and cutting-edge research |
| Eleven Flash v2.5 | ElevenLabs | $33.00 | $0 | 5,000 | Budget | Cheapest TTS (~$0.033/1K chars) |
| Luminous Supreme | Aleph Alpha | $35.00 | $35.00 | 2,048 | Premium | Premium multilingual completions on EU-sovereign infrastructure |
| Eleven Turbo v2.5 | ElevenLabs | $50.00 | $0 | 5,000 | Mid-tier | Low-latency TTS (~$0.05/1K chars) |
| MiniMax Speech-2.5 HD | MiniMax | $60.00 | $0 | 5,000 | Flagship | Multilingual expressive TTS (~$0.06/1K chars) — 30+ languages |
| GPT-4.5 | OpenAI | $75.00 | $150.00 | 128,000 | Premium | Original GPT-4.5 research preview — broad knowledge and writing |
| Eleven Multilingual v2 | ElevenLabs | $180.00 | $0 | 5,000 | Premium | TTS billed per character (~$0.18/1K chars) |