AI Model Pricing Comparison — API Costs for 49+ Models
Compare costs across all models. Calculate your monthly spend based on usage.
Estimate your cost
| Model | Provider | Cost Visualization | ||||
|---|---|---|---|---|---|---|
| GPT-5 Nano budget | OpenAI | $0.05 | $0.40 | 400K | 250 tok/s | |
| Nova Lite budget | Amazon | $0.06 | $0.24 | 300K | 150 tok/s | |
| Gemini 2.0 Flash budget | $0.10 | $0.40 | 1.0M | 220 tok/s | ||
| GPT-4.1 Nano budget | OpenAI | $0.10 | $0.40 | 1.0M | 200 tok/s | |
| Llama 3.3 70B open source | Meta | $0.10 | $0.30 | 131K | 90 tok/s | |
| Qwen 3 32B open source | Qwen | $0.10 | $0.30 | 131K | 120 tok/s | |
| Mistral Small 3.2 budget | Mistral | $0.10 | $0.30 | 131K | 160 tok/s | |
| Gemini 2.5 Flash Lite budget | $0.10 | $0.40 | 1.0M | 240 tok/s | ||
| Qwen 2.5 72B open source | Qwen | $0.12 | $0.39 | 131K | 80 tok/s | |
| DeepSeek V3.1 open source | DeepSeek | $0.14 | $0.28 | 131K | 65 tok/s | |
| DeepSeek V3.2 open source | DeepSeek | $0.14 | $0.28 | 131K | 70 tok/s | |
| Yi Lightning mid | 01.AI | $0.14 | $0.14 | 16K | — | |
| GPT-4o Mini budget | OpenAI | $0.15 | $0.60 | 128K | 200 tok/s | |
| Command R mid | Cohere | $0.15 | $0.60 | 128K | 90 tok/s | |
| Llama 4 Scout open source | Meta | $0.17 | $0.50 | 10.0M | 140 tok/s | |
| Grok 4 Fast mid | xAI | $0.20 | $0.50 | 256K | 110 tok/s | |
| Grok 4.1 Fast mid | xAI | $0.20 | $0.50 | 256K | 120 tok/s | |
| Qwen 3 235B open source | Qwen | $0.20 | $0.80 | 131K | 60 tok/s | |
| Jamba 1.5 Mini mid | AI21 Labs | $0.20 | $0.40 | 256K | 35 tok/s | |
| Reka Flash 3 mid | Reka AI | $0.20 | $0.80 | 128K | 136 tok/s | |
| Qwen 3 Coder open source | Qwen | $0.20 | $0.80 | 131K | 80 tok/s | |
| GPT-5 Mini mid | OpenAI | $0.25 | $2.0 | 400K | 180 tok/s | |
| Llama 4 Maverick open source | Meta | $0.25 | $0.80 | 1.0M | 100 tok/s | |
| GPT-5.1 Codex Mini mid | OpenAI | $0.25 | $2.0 | 400K | 170 tok/s | |
| Gemini 2.5 Flash mid | $0.30 | $2.5 | 1.0M | 251 tok/s | ||
| Grok 3 Mini budget | xAI | $0.30 | $0.50 | 131K | 130 tok/s | |
| Codestral mid | Mistral | $0.30 | $0.90 | 262K | 90 tok/s | |
| GPT-4.1 Mini budget | OpenAI | $0.40 | $1.6 | 1.0M | 160 tok/s | |
| Mistral Medium 3.1 mid | Mistral | $0.40 | $2.0 | 131K | 110 tok/s | |
| Qwen 3 Max open source | Qwen | $0.46 | $1.8 | 131K | 55 tok/s | |
| Gemini 3 Flash mid | $0.50 | $3.0 | 1.0M | 200 tok/s | ||
| Mistral Large 25.12 flagship | Mistral | $0.50 | $1.5 | 131K | 70 tok/s | |
| DeepSeek R1 open source | DeepSeek | $0.55 | $2.2 | 131K | 35 tok/s | |
| DeepSeek R1 0528 open source | DeepSeek | $0.55 | $2.2 | 131K | 40 tok/s | |
| Qwen 3.5 397B open source | Qwen | $0.60 | $3.6 | 131K | 45 tok/s | |
| Claude 3.5 Haiku budget | Anthropic | $0.80 | $4.0 | 200K | 170 tok/s | |
| Nova Pro mid | Amazon | $0.80 | $3.2 | 300K | 100 tok/s | |
| Claude Haiku 4.5 budget | Anthropic | $1.0 | $5.0 | 200K | 180 tok/s | |
| Sonar mid | Perplexity | $1.0 | $1.0 | 128K | — | |
| o3 Mini mid | OpenAI | $1.1 | $4.4 | 200K | 100 tok/s | |
| o4 Mini mid | OpenAI | $1.1 | $4.4 | 200K | 120 tok/s | |
| GPT-5.1 flagship | OpenAI | $1.3 | $10.0 | 400K | 95 tok/s | |
| GPT-5 flagship | OpenAI | $1.3 | $10.0 | 400K | 100 tok/s | |
| Gemini 2.5 Pro flagship | $1.3 | $10.0 | 1.0M | 55 tok/s | ||
| GPT-5.1 Codex flagship | OpenAI | $1.3 | $10.0 | 400K | 85 tok/s | |
| GPT-5.2 flagship | OpenAI | $1.8 | $14.0 | 400K | 90 tok/s | |
| o3 flagship | OpenAI | $2.0 | $8.0 | 200K | 40 tok/s | |
| Gemini 3.1 Pro flagship | $2.0 | $12.0 | 1.0M | 65 tok/s | ||
| Gemini 3 Pro flagship | $2.0 | $12.0 | 1.0M | 60 tok/s | ||
| Pixtral Large flagship | Mistral | $2.0 | $6.0 | 131K | 60 tok/s | |
| Sonar Reasoning Pro flagship | Perplexity | $2.0 | $8.0 | 128K | 22 tok/s | |
| Jamba 1.5 Large flagship | AI21 Labs | $2.0 | $8.0 | 256K | 19 tok/s | |
| Grok 2 mid | xAI | $2.0 | $10.0 | 131K | 80 tok/s | |
| GPT-4.1 mid | OpenAI | $2.0 | $8.0 | 1.0M | 70 tok/s | |
| GPT-4o mid | OpenAI | $2.5 | $10.0 | 128K | 143 tok/s | |
| Command R+ flagship | Cohere | $2.5 | $10.0 | 128K | 60 tok/s | |
| Command A flagship | Cohere | $2.5 | $10.0 | 256K | 70 tok/s | |
| o1 Mini mid | OpenAI | $3.0 | $12.0 | 128K | 80 tok/s | |
| Claude Sonnet 4.6 flagship | Anthropic | $3.0 | $15.0 | 200K | 57 tok/s | |
| Claude Sonnet 4 mid | Anthropic | $3.0 | $15.0 | 200K | 75 tok/s | |
| Claude 3.7 Sonnet mid | Anthropic | $3.0 | $15.0 | 200K | 70 tok/s | |
| Claude Sonnet 4.5 flagship | Anthropic | $3.0 | $15.0 | 200K | 67 tok/s | |
| Claude 3.5 Sonnet mid | Anthropic | $3.0 | $15.0 | 200K | 70 tok/s | |
| Grok 3 mid | xAI | $3.0 | $15.0 | 1.0M | 65 tok/s | |
| Grok 4 flagship | xAI | $3.0 | $15.0 | 256K | 55 tok/s | |
| Sonar Pro flagship | Perplexity | $3.0 | $15.0 | 200K | 50 tok/s | |
| Claude Opus 4.6 flagship | Anthropic | $5.0 | $25.0 | 200K | 68 tok/s | |
| Claude Opus 4.5 flagship | Anthropic | $5.0 | $25.0 | 200K | 50 tok/s | |
| o1 flagship | OpenAI | $15.0 | $60.0 | 200K | 35 tok/s | |
| Claude Opus 4 flagship | Anthropic | $15.0 | $75.0 | 200K | 50 tok/s | |
| Claude 3 Opus flagship | Anthropic | $15.0 | $75.0 | 200K | 25 tok/s | |
| o3 Pro flagship | OpenAI | $20.0 | $80.0 | 200K | 25 tok/s | |
| GPT-4.5 flagship | OpenAI | $75.0 | $150.0 | 128K | 60 tok/s | |
| GLM-5 flagship | Zhipu AI | — | — | 200K | 55 tok/s | — |
| GLM-4.6V mid | Zhipu AI | — | — | 128K | 68 tok/s | — |
| Kimi K2.5 flagship | Moonshot AI | — | — | 256K | 45 tok/s | — |
| MiniMax M2.5 flagship | MiniMax | — | — | 205K | 59 tok/s | — |
| MiMo V2 Flash budget | Xiaomi | — | — | 256K | 155 tok/s | — |
| Doubao Seed 2.0 flagship | ByteDance | — | — | 128K | — | — |
| Phi-4 open source | Microsoft | — | — | 16K | 93 tok/s | — |
| Phi-4 Mini open source | Microsoft | — | — | 128K | 150 tok/s | — |
| Phi-4 Reasoning Plus open source | Microsoft | — | — | 32K | — | — |
| Nemotron 3 Nano open source | NVIDIA | — | — | 1.0M | 76 tok/s | — |
| Gemma 3 27B open source | — | — | 128K | 58 tok/s | — | |
| Gemma 3 12B open source | — | — | 131K | 90 tok/s | — | |
| Gemma 3 4B open source | — | — | 96K | — | — | |
| Magistral Medium 1.2 mid | Mistral | — | — | 128K | 29 tok/s | — |
| Mistral Large 3 open source | Mistral | — | — | 256K | — | — |
| Ministral 3 14B open source | Mistral | — | — | 256K | 83 tok/s | — |
| Ministral 3 8B open source | Mistral | — | — | 256K | 172 tok/s | — |
| Qwen 3 Next 80B open source | Qwen | — | — | 262K | 138 tok/s | — |
| Qwen 3 Coder 480B open source | Qwen | — | — | 262K | 60 tok/s | — |
| Qwen 3 VL 235B open source | Qwen | — | — | 262K | 46 tok/s | — |
| GPT-OSS 20B open source | OpenAI | — | — | 131K | 312 tok/s | — |
| GPT-OSS 120B open source | OpenAI | — | — | 131K | 339 tok/s | — |
| Nova 2.0 Lite budget | Amazon | — | — | 1.0M | 221 tok/s | — |
| K-EXAONE flagship | LG AI Research | — | — | 256K | 62 tok/s | — |
| Ring Flash 2.0 open source | InclusionAI | — | — | 128K | 83 tok/s | — |
| Step 2.5 Flash mid | StepFun | — | — | 128K | 67 tok/s | — |
| Sora 2 flagship | OpenAI | — | — | — | — | — |
| Veo 3.1 flagship | — | — | — | — | — | |
| Veo 3 flagship | — | — | — | — | — | |
| Kling 2.5 Turbo flagship | Kuaishou | — | — | — | — | — |
| DALL-E 3 flagship | OpenAI | — | — | — | — | — |
| Midjourney v7 flagship | Midjourney | — | — | — | — | — |
| Midjourney v6.1 flagship | Midjourney | — | — | — | — | — |
| Stable Diffusion 3.5 open source | Stability AI | — | — | — | — | — |
| Flux 1.1 Pro flagship | Black Forest Labs | — | — | — | — | — |
| Flux 1.0 Dev open source | Black Forest Labs | — | — | — | — | — |
| Imagen 4 flagship | — | — | — | — | — | |
| Ideogram 3.0 flagship | Ideogram | — | — | — | — | — |
| Seedream 4.5 flagship | ByteDance | — | — | — | — | — |
| ERNIE 4.5 flagship | Baidu | — | — | 128K | — | — |
| ERNIE X1 flagship | Baidu | — | — | 128K | — | — |
| Hermes 4 70B open source | Nous Research | — | — | 128K | — | — |
| Luma Ray 3 flagship | Luma AI | — | — | — | — | — |
| Runway Gen-4 flagship | Runway | — | — | — | — | — |
| Pika 2.5 flagship | Pika | — | — | — | — | — |
Pricing sourced from OpenRouter API. Prices may vary by provider and tier. Last updated daily.
AI Pricing FAQ
How much does the OpenAI API cost?
OpenAI API pricing varies by model. GPT-5.2 costs $1.75/1M input tokens and $14/1M output tokens. GPT-5 costs $1.25/$10. Budget options like GPT-5 Nano cost just $0.05/$0.40, and GPT-4.1 Nano costs $0.10/$0.40. Reasoning models like o3 cost $2.00/$8.00 per 1M tokens. Prices are current as of 2026.
Which AI API is cheapest?
The cheapest AI APIs by input token cost are: GPT-5 Nano ($0.05/1M), Nova Lite ($0.06/1M), GPT-4.1 Nano and Gemini 2.5 Flash Lite ($0.10/1M each), and DeepSeek V3.2 ($0.14/1M). For the best quality-to-price ratio, DeepSeek V3.2 and Qwen models offer exceptional value with near-flagship performance at budget prices.
How do AI model pricing tiers work?
AI model pricing is based on token usage — you pay per million tokens processed. Input tokens (your prompts) and output tokens (model responses) are priced separately, with output tokens typically costing 2-8x more. Flagship models (GPT-5.2, Claude Opus 4.6) cost $1-5/1M input, mid-tier models cost $0.20-1.00/1M, and budget models cost $0.05-0.15/1M input tokens.
What are input vs output tokens?
Input tokens are the text you send to the AI model (your prompts, context, system instructions). Output tokens are the text the model generates in response. A token is roughly 3/4 of a word in English. Output tokens are more expensive because they require more compute to generate. For example, GPT-5.2 charges $1.75/1M input tokens but $14/1M output tokens — 8x more for output.