Tool
AI API cost calculator
Enter your usage for a month and compare the estimated API cost of every model with verified current prices. Prices stay in their own currency; nothing is converted.
Not sure how many tokens you use? AI token pricing explained shows what a token is and how input, output and cached tokens are billed, and how to reduce AI API costs covers caching, batching and other ways to lower the bill. Every price is listed in the AI pricing table.
Example: 1,000,000 input and 250,000 output tokens per month
Per-request size unknown: enter the number of requests to check context and output limits.
| # | Model | Total | Breakdown | Prices as of |
|---|---|---|---|---|
| 1 | Amazon Nova Micro Amazon | $0.07 |
|
2026-10-25 Amazon Bedrock Pricing |
| 2 | Command R7B Cohere | $0.075 |
|
2026-10-25 Pricing | Secure and Scalable Enterprise AI | Cohere |
| 3 | Amazon Nova Lite Amazon | $0.12 |
|
2026-10-25 Amazon Bedrock Pricing |
| 4 | Ministral 3 3B Mistral AI | $0.125 |
|
2026-10-25 API pricing | Mistral Docs |
| 5 | Ministral 3 8B Mistral AI | $0.1875 |
|
2026-10-25 API pricing | Mistral Docs |
| 6 | Claude Haiku 5.5 Anthropic | $0.225 |
|
2026-10-25 Pricing - Claude Platform Docs |
| 7 | GPT-6 Luna OpenAI | $0.225 |
|
2026-10-25 Pricing | OpenAI API |
| 8 | Ministral 3 14B Mistral AI | $0.25 |
|
2026-10-25 API pricing | Mistral Docs |
| 9 | Qwen3.8-Flash Alibaba Cloud (Qwen) | $0.2675 |
|
2026-10-25 qwen3.8-flash Model Info - Alibaba Cloud Model Studio |
| 10 | Qwen3.8-Omni-Flash Alibaba Cloud (Qwen) | $0.2675 |
|
2026-10-25 Alibaba Cloud Model Studio model pricing |
| 11 | Command R (08-2024) Cohere | $0.30 |
|
2026-10-25 Cohere's Command R Model | Cohere |
| 12 | Mistral Small 4 Mistral AI | $0.30 |
|
2026-10-25 API pricing | Mistral Docs |
| 13 | DeepSeek-V4.1-Flash DeepSeek | $0.60 |
|
2026-10-25 Models & Pricing | DeepSeek API Docs |
| 14 | Gemini 3.1 Flash-Lite Google | $0.625 |
|
2026-10-25 Gemini Developer API pricing |
| 15 | Nano Banana 2 Lite Google | $0.625 |
|
2026-10-25 Gemini Developer API pricing |
| 16 | Qwen3.7-Plus Alibaba Cloud (Qwen) | $0.80 |
|
2026-10-25 qwen3.7-plus Model Info - Alibaba Cloud Model Studio |
| 17 | Mistral Large 3 Mistral AI | $0.875 |
|
2026-10-25 API pricing | Mistral Docs |
| 18 | Amazon Nova 2 Lite Amazon | $0.925 |
|
2026-10-25 Amazon Bedrock Pricing |
| 19 | Gemini 3.5 Flash-Lite Google | $0.925 |
|
2026-10-25 Gemini Developer API pricing |
| 20 | Amazon Nova 2 Sonic Amazon | $1.0175 |
|
2026-10-25 Amazon Bedrock Pricing |
| 21 | Amazon Nova 2.5 Sonic Amazon | $1.0175 |
|
2026-10-25 Amazon Bedrock Pricing |
| 22 | Qwen3.8-27B Alibaba Cloud (Qwen) | $1.25 |
|
2026-10-25 qwen3.8-27b Model Info - Alibaba Cloud Model Studio |
| 23 | Grok Build 0.1 xAI | $1.50 |
|
2026-10-25 API Pricing |
| 24 | Amazon Nova Pro Amazon | $1.60 |
|
2026-10-25 Amazon Bedrock Pricing |
| 25 | Gemini 3.6 Flash Google | $1.6875 |
|
2026-10-25 Gemini Developer API pricing |
| 26 | Gemini 3.8 Flash Google | $1.6875 |
|
2026-10-25 Gemini Developer API pricing |
| 27 | Grok 4.20 (Non-Reasoning) xAI | $1.875 |
|
2026-10-25 API Pricing |
| 28 | Grok 4.20 (Reasoning) xAI | $1.875 |
|
2026-10-25 API Pricing |
| 29 | Grok 4.20 Multi-Agent (Beta) xAI | $1.875 |
|
2026-10-25 API Pricing |
| 30 | Grok 4.3 xAI | $1.875 |
|
2026-10-25 API Pricing |
| 31 | DeepSeek-V4-Pro DeepSeek | $2.31 |
|
2026-10-25 Models & Pricing | DeepSeek API Docs |
| 32 | Muse Spark 1.1 Meta | $2.3125 |
|
2026-10-25 Pricing and rate limits - Meta Model API |
| 33 | Muse Spark 1.2 Meta | $2.3125 |
|
2026-10-25 Pricing and rate limits - Meta Model API |
| 34 | Muse Spark 1.3 Meta | $2.3125 |
|
2026-10-25 Pricing and rate limits - Meta Model API |
| 35 | Mistral Large 4 Mistral AI | $2.405 |
|
2026-10-25 API pricing | Mistral Docs |
| 36 | Mistral Medium 3.5 Mistral AI | $3.375 |
|
2026-10-25 API pricing | Mistral Docs |
| 37 | Nano Banana 2.1 Google | $3.375 |
|
2026-10-25 Gemini Developer API pricing |
| 38 | Grok 4.5 xAI | $3.50 |
|
2026-10-25 API Pricing |
| 39 | Grok 4.6 xAI | $3.50 |
|
2026-10-25 API Pricing |
| 40 | Grok 4.7 xAI | $3.50 |
|
2026-10-25 API Pricing |
| 41 | Qwen3.8-2.4T-A95B Alibaba Cloud (Qwen) | $3.50 |
|
2026-10-25 qwen3.8-2.4t-a95b Model Info - Alibaba Cloud Model Studio |
| 42 | Qwen3.8-Max Alibaba Cloud (Qwen) | $3.50 |
|
2026-10-25 qwen3.8-max Model Info - Alibaba Cloud Model Studio |
| 43 | Claude Sonnet 5.5 Anthropic | $4.50 |
|
2026-10-25 Pricing - Claude Platform Docs |
| 44 | Gemini 4 Argon Google | $4.50 |
|
2026-10-25 Gemini Developer API pricing |
| 45 | GPT-6 Sol OpenAI | $4.50 |
|
2026-10-25 Pricing | OpenAI API |
| 46 | GPT-6.1 Sol OpenAI | $4.50 |
|
2026-10-25 Pricing | OpenAI API |
| 47 | Command A Cohere | $5.00 |
|
2026-10-25 Command A | Cohere |
| 48 | Gemini 3.1 Pro Google | $5.00 |
|
2026-10-25 Gemini Developer API pricing |
| 49 | Nano Banana Pro Google | $5.00 |
|
2026-10-25 Gemini Developer API pricing |
| 50 | Claude Sonnet 4.5 Anthropic | $6.75 |
|
2026-10-25 Pricing - Claude Platform Docs |
| 51 | Claude Opus 5.5 Anthropic | $9.00 |
|
2026-10-25 Pricing - Claude Platform Docs |
| 52 | Muse Wonder Meta | $9.00 |
|
2026-10-25 Pricing and rate limits - Meta Model API |
| 53 | Claude Fable 5.1 Anthropic | $22.50 |
|
2026-10-25 Pricing - Claude Platform Docs |
| 54 | GPT-6 Astra OpenAI | $22.50 |
|
2026-10-25 Pricing | OpenAI API |
Not estimated (10)
- Command A Reasoning: no input/output price listed
- Command A Vision: no input/output price listed
- Command A+: no input/output price listed
- Gemma 4 31B: no input/output price listed
- gpt-oss-120b: no input/output price listed
- gpt-oss-20b: no input/output price listed
- Llama 4 Maverick: no input/output price listed
- Llama 4 Scout: no input/output price listed
- Muse Image: no input/output price listed
- North Mini Code: no input/output price listed
How the estimate is calculated
Each model's cost is the sum of fresh input tokens, cached input tokens and output tokens multiplied by its current published price for the matching dimension, rounded to the nearest millionth of a currency unit. With batch processing the batch input and output prices are used; models without batch prices are listed as excluded. Cached tokens use the model's cached-input price, or its input price when no cache price is listed (flagged).
Only models that produce text are compared; image-only and speech-only models are priced in their own dimensions on the pricing page.
Your numbers are totals for the period. Enter the number of requests to check context-window and output limits against the average request size and to pick the right price tier; without it the base tier is assumed and flagged.