AI Model API Pricing

Published input and output prices per million tokens, context windows and access notes for the models tracked by Frontrumors. Dates and restrictions remain specific to each source.

Workload cost calculator

Small app · 12M input + 3M output / month100 requests/day × 30 days · 4,000 in / 1,000 out per request · Cache off
Workload cost assumptions
Editable examples

Illustrative Small app example; all values are editable.

Start with
Expected request usage
Prompt and context tokens.
Generated response tokens.
Daily average, including fractions.
Adjustable, 1–366 whole days.
Context and caching options
Optional caching scenario

Unpriced cache writes or storage make the estimate partial and exclude it from ranking.

Monthly comparison

Example: 4,000 input and 1,000 output tokens per request, 100 requests a day for 30 days.

Ready to compare when the calculator loads.

Token estimates exclude taxes, images, audio, tools and account discounts. Partial or unavailable models are never ranked.

    Published token rates

    Base rates are in USD per million tokens. Context tiers, cache billing and access restrictions are shown with the calculator assumptions.

    API prices in US dollars per million tokens
    ModelInputOutputContextLabSource and check date
    Gemini 4 Carbon$2$10Unconfirmed in Google announcementGoogleGoogle Carbon announcement and pricing Last checked:
    Claude Opus 5.5$4$201MAnthropicAnthropic pricing Last checked:
    GPT-6 Astra$10$501.05MOpenAIOpenAI API pricing Last checked:
    GPT-6.1 Sol$2$101.05MOpenAIOpenAI API pricing Last checked:
    Claude Sonnet 5.5$2$101MAnthropicAnthropic pricing Last checked:
    GPT-6.1 Terra$1$51.05MOpenAIOpenAI API pricing Last checked:
    Gemini 4 Argon$2$101MGoogleGoogle Gemini API pricing Last checked:
    Muse Wonder Full (preview)Not verifiedNot verified1MMetaMeta Muse Wonder models and API availability Last checked:
    Claude Fable 5.1$10$501MAnthropicAnthropic pricing Last checked:
    Kimi K3.1 Pro$2$101MMoonshotKimi API pricing Last checked:
    Claude Haiku 5.5$0.10$0.501MAnthropicAnthropic pricing Last checked:
    Grok 4.7$2$6500KxAIxAI API pricing Last checked:
    Qwen3.8-Max$2$61MAlibabaAlibaba Qwen API pricing Last checked:
    DeepSeek V4.1-Flash-1017$0.30$1.201MDeepSeekDeepSeek API pricing and model version Last checked:
    Mistral Large 4 (preview)$0.70$2.101MMistralMistral inference pricing Last checked:

    Access, pricing and verification notes

    Gemini 4 Carbon

    Restricted access; announced introductory $2/$10, then $4/$20 with no dated end. AA cost-per-task uses standard pricing

    Google Carbon announcement and pricing Last checked:

    Claude Opus 5.5

    Anthropic list price

    Anthropic pricing Last checked:

    GPT-6 Astra

    Limited orgs first; 'GPT-6 Pro' in ChatGPT Pro

    OpenAI API pricing Last checked:

    GPT-6.1 Sol

    OpenAI list price

    OpenAI API pricing Last checked:

    Claude Sonnet 5.5

    Anthropic list price

    Anthropic pricing Last checked:

    GPT-6.1 Terra

    OpenAI list price; launched 26 Oct 2026

    OpenAI API pricing Last checked:

    Gemini 4 Argon

    Doubles to $4/$20 on 1 Dec 2026; first Gemini 4 model, Stable GA 13 Oct

    Google Gemini API pricing Last checked:

    Muse Wonder Full (preview)

    Muse Wonder Full is a Muse Code subscription model with no published per-token API price. The adjacent Muse Wonder API row is a different model.

    Meta Muse Wonder models and API availability Last checked:

    Claude Fable 5.1

    Mythos-class flagship, GA day one (1 Sep)

    Anthropic pricing Last checked:

    Kimi K3.1 Pro

    Kimi API homepage: standard $2/$10, Highspeed $4/$20; checked 1 Nov. Weights promised by 4 Nov, not yet verified

    Kimi API pricing Last checked:

    Claude Haiku 5.5

    Anthropic list price up to 100K-token prompts; $0.50/$2.50 above that

    Anthropic pricing Last checked:

    Grok 4.7

    Same price as Grok 4.6; higher rates above 200K tokens of context

    xAI API pricing Last checked:

    Qwen3.8-Max

    Verified Alibaba International (Singapore) / US (Virginia) pricing $2/$6; Global scope is $1.65/$4.951. Implicit-cache read $0.25 in the selected region.

    Alibaba Qwen API pricing Last checked:

    DeepSeek V4.1-Flash-1017

    Peak-rate default $0.30/$1.20; off-peak $0.15/$0.60, with UTC windows and holiday conditions. Verified first-party pricing; the API serves V4.1-Flash-1017.

    DeepSeek API pricing and model version Last checked:

    Mistral Large 4 (preview)

    Mistral docs: Public Preview, 1M context, $0.70/$2.10 and $0.07 cached input; checked 2 Nov. Weights released under the Mistral Ecosystem Open-Weight License; separate Model-as-a-Service terms

    Mistral inference pricing Last checked:

    Pricing alerts

    Price sources

    A listed price does not imply general availability. Confirm current terms with the lab before committing. Editorial recommendations are available in Reviews.