AI model pricing
Compare verified input and output token prices across leading AI providers.
Lowest input: $0.035 (Amazon Nova Micro) Lowest output: $0.14 (Amazon Nova Micro) Largest context: 1050K (GPT-5.4 / GPT-5.5)
37 pricing records
| Model | Provider | InputUSD / 1M tokens | OutputUSD / 1M tokens | Context |
|---|---|---|---|---|
| Amazon Nova Micro | Amazon | $0.035 | $0.14 | 128K |
| Command R7B | Cohere | $0.0375 | $0.15 | 128K |
| Amazon Nova Lite | Amazon | $0.06 | $0.24 | 300K |
| Granite 4H Small⚠ | IBM | $0.0636 | $0.265 | 131K |
| Muse Spark 1.3 (Contributor)⚠ | Meta | $0.1 | $0.2 | 1M |
| Mistral Small 4⚠ | Mistral | $0.15 | $0.6 | 256K |
| Qwen3.8 Flash | Qwen | $0.15 | $0.47 | 992K |
| Command R | Cohere | $0.15 | $0.6 | 128K |
| Jamba Mini⚠ | AI21 | $0.2 | $0.4 | 256K |
| GLM-4.5-Air | Z.ai | $0.2 | $1.1 | 128K |
| DeepSeek V4-Flash⚠ | DeepSeek | $0.22 | $0.66 | 1M |
| Gemini 2.5 Flash | $0.3 | $2.5 | 1.05M | |
| MiniMax-M3⚠ | MiniMax | $0.3† | $1.2 | 1M |
| Amazon Nova 2 Lite | Amazon | $0.33 | $2.75 | 1M |
| Qwen3.7 Plus | Qwen | $0.4† | $1.6 | 992K |
| Mistral Large 3⚠ | Mistral | $0.5 | $1.5 | 256K |
| GLM-4.6 | Z.ai | $0.6 | $2.2 | 200K |
| DeepSeek V4-Pro⚠ | DeepSeek | $0.66 | $1.98 | 1M |
| GPT-5.4 Mini | OpenAI | $0.75 | $4.5 | 400K |
| Amazon Nova Pro | Amazon | $0.8 | $3.2 | 300K |
| Kimi K2.6 | Kimi | $0.95 | $4 | 262K |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | 200K |
| Grok Build 0.1 | xAI | $1† | $2 | 256K |
| Gemini 2.5 Pro | $1.25† | $10 | 1.05M | |
| Grok 4.3 | xAI | $1.25† | $2.5 | 1M |
| Muse Spark 1.3⚠ | Meta | $1.25 | $4.25 | 1M |
| GLM-5.3 | Z.ai | $1.4 | $4.4 | 1M |
| Claude Sonnet 5⚠ | Anthropic | $2 | $10 | 1M |
| Qwen3.8 Max | Qwen | $2 | $6 | 992K |
| Grok 4.6 | xAI | $2† | $6 | 500K |
| Jamba Large⚠ | AI21 | $2 | $8 | 256K |
| GPT-5.4 | OpenAI | $2.5† | $15 | 1.05M |
| Amazon Nova Premier | Amazon | $2.5 | $12.5 | 1M |
| Kimi K3 | Kimi | $3 | $15 | 1.05M |
| Claude Opus 5 | Anthropic | $5 | $25 | 1M |
| GPT-5.5 | OpenAI | $5† | $30 | 1.05M |
| Claude Fable 5.1⚠ | Anthropic | $10 | $50 | 1M |
- ⚠ has an important pricing note (hover or see the model row)
- † has a long-context surcharge above a token threshold (hover for detail)
Pricing by provider
OpenAI
OpenAI prices its GPT-5 models per million input and output tokens. Some models apply a higher rate once a single request's input tokens exceed a documented threshold.
- GPT-5.4 Mini: $0.75 / 1M input tokens, $4.50 / 1M output tokens
- GPT-5.5: $5.00 / 1M input tokens, $30.00 / 1M output tokens — above 272K input tokens per request: $10.00 / $45.00
Anthropic / Claude
Anthropic prices Claude models per million input and output tokens. Claude Sonnet 5's pricing was re-verified directly against Anthropic's official documentation on 2026-09-06.
- Claude Haiku 4.5: $1.00 / 1M input tokens, $5.00 / 1M output tokens
- Claude Sonnet 5: $2.00 / 1M input tokens, $10.00 / 1M output tokens
- Claude Opus 5: $5.00 / 1M input tokens, $25.00 / 1M output tokens
Google / Gemini
Google prices Gemini models per million input and output tokens. Gemini 2.5 Pro applies a higher rate once a single request's input tokens exceed a documented threshold; Gemini 2.5 Flash does not.
- Gemini 2.5 Flash: $0.30 / 1M input tokens, $2.50 / 1M output tokens
- Gemini 2.5 Pro: $1.25 / 1M input tokens, $10.00 / 1M output tokens — above 200K input tokens per request: $2.50 / $15.00
Mistral
Mistral prices its models per million input and output tokens. Mistral Large 3 is priced higher than Mistral Small 4 in T9 Token Budget's current pricing data.
- Mistral Small 4: $0.15 / 1M input tokens, $0.60 / 1M output tokens
- Mistral Large 3: $0.50 / 1M input tokens, $1.50 / 1M output tokens
DeepSeek
DeepSeek prices its V4 models per million input and output tokens. Provider pricing can change, so verify current rates with DeepSeek before budgeting.
- DeepSeek V4-Flash: $0.22 / 1M input tokens, $0.66 / 1M output tokens
- DeepSeek V4-Pro: $0.66 / 1M input tokens, $1.98 / 1M output tokens
Kimi
Kimi (Moonshot AI) prices its models per million input and output tokens on its international, USD-denominated API. Its mainland-China platform publishes an independent CNY price list, not shown here.
- Kimi K2.6: $0.95 / 1M input tokens, $4.00 / 1M output tokens
- Kimi K3: $3.00 / 1M input tokens, $15.00 / 1M output tokens
Qwen
Qwen (Alibaba Cloud) prices its models per million input and output tokens on its international API. Qwen3.7 Plus applies a higher rate once a single request's input tokens exceed a documented threshold.
- Qwen3.8 Flash: $0.15 / 1M input tokens, $0.47 / 1M output tokens
- Qwen3.7 Plus: $0.40 / 1M input tokens, $1.60 / 1M output tokens — above 256K input tokens per request: $1.20 / $4.80
- Qwen3.8 Max: $2.00 / 1M input tokens, $6.00 / 1M output tokens
Z.ai
Z.ai prices its GLM models per million input and output tokens on its international, USD-denominated API. GLM-5.3 is its current flagship; GLM-4.6 and GLM-4.5-Air remain separately available at their own rates.
- GLM-4.5-Air: $0.20 / 1M input tokens, $1.1 / 1M output tokens
- GLM-4.6: $0.60 / 1M input tokens, $2.2 / 1M output tokens
- GLM-5.3: $1.40 / 1M input tokens, $4.4 / 1M output tokens
MiniMax
MiniMax prices its models per million input and output tokens on its international, USD-denominated API. MiniMax-M3's listed rate reflects the provider's own stated permanent discount off its list price.
- MiniMax-M3: $0.30 / 1M input tokens, $1.20 / 1M output tokens — above 512K input tokens per request: $0.60 / $2.40
xAI
xAI prices its Grok models per million input and output tokens. Grok 4.3 and Grok 4.6 apply a higher rate once a single request's input tokens reach a documented threshold.
- Grok Build 0.1: $1.00 / 1M input tokens, $2.00 / 1M output tokens — above 200K input tokens per request: $2.00 / $4.00
- Grok 4.3: $1.25 / 1M input tokens, $2.50 / 1M output tokens — above 200K input tokens per request: $2.50 / $5.00
- Grok 4.6: $2.00 / 1M input tokens, $6.00 / 1M output tokens — above 200K input tokens per request: $4.00 / $12.00
Cohere
Cohere prices its Command R models per million input and output tokens. Its newest flagship, Command A+, is distributed as free open-weights rather than a metered API and is not included in this pricing data.
- Command R7B: $0.0375 / 1M input tokens, $0.15 / 1M output tokens
- Command R: $0.15 / 1M input tokens, $0.60 / 1M output tokens
How to read AI pricing
Input tokens
What you send to the model — your prompt, conversation history, or document text. Priced separately from output, usually at a lower rate.
Output tokens
What the model generates in response. Output is commonly the more expensive side — often 3–10× the input price — because generating text costs more compute than reading it.
Context window
How much text a model can hold in one request, in tokens. A larger window lets you include more history or documents, but filling it means paying for more input on every call.
Whose prices are these?
The prices in the table above belong to the AI providers (OpenAI, Anthropic, Google, Mistral, and DeepSeek — every provider currently priced in the calculator), not to T9. T9 does not sell, resell, or provide access to any model — it only uses published prices to produce an estimate.
Prices change — always re-verify
AI providers update prices frequently. The figures above were manually verified on 2026-09-08, but pricing should always be rechecked on the provider's official pricing page before any budget decision. T9 uses pricing data only for estimation, and never as a guarantee of what you will be charged.
See pricing in context
Beyond comparing raw prices, T9 lets you apply these current rates to your own use case. Open the chatbot, meeting summary, document summary, or document Q&A calculator and switch models to see how the same workload changes cost, or read the models overview and the methodology.