← the bench index · API pricing
LLM API pricing comparison
The #1 model on the bench index, Claude Opus 5.5, costs $4 / $20 per million tokens (input / output). The cheapest input is $0.10 (Claude Haiku 5.5 and GPT-6 Luna); the priciest output is $50 (GPT-6 Astra, Fable 5.1 and Fable 5).
What 40 live models cost on their makers' own APIs, in US dollars per million tokens, from 12 labs, as of November 8, 2026. Your own workload: the LLM API cost calculator.
Get the next launch alert, free →
By lab: Anthropic · OpenAI · Google · Meta · xAI · Zhipu AI · Alibaba · MiniMax · Moonshot · Xiaomi · prompt caching prices · context windows · knowledge cutoffs · model retirements · pricing terms explained · price against quality, charted
Models on the bench index
| Model | Input | Output | Cached input | Context | AW Index |
|---|---|---|---|---|---|
Claude Opus 5.5Anthropic · claude-opus-5-5 | $4 | $20 | $0.20 | 1M | 98.1 |
GPT-6 AstraOpenAI · gpt-6-astra | $10† | $50† | $1 | 1.05M | 92.6 |
Fable 5.1Anthropic · claude-fable-5-1 | $10 | $50 | $0.25 | 1M | 91.7 |
GPT-6.1 SolOpenAI · gpt-6.1-sol | $2† | $10† | $0.10 | 1.05M | 88.4 |
Claude Sonnet 5.5Anthropic · claude-sonnet-5-5 | $2 | $10 | $0.10 | 1M | 84.6 |
Gemini 4 ArgonGoogle · gemini-4-argon | $2 | $10 | $0.10 | 1.05M | 80 |
GPT-6.1 TerraOpenAI · gpt-6.1-terra | $1† | $5† | $0.05 | 1.05M | 77.6 |
Claude Opus 5Anthropic · claude-opus-5 | $5 | $25 | $0.50 | 1M | 74.1 |
Kimi K3.1 ProMoonshot · kimi-k3.1-pro | $2 | $10 | $0.20 | 1.05M | 66.7 |
GPT-6 SolOpenAI · gpt-6-sol | $2† | $10† | $0.20 | 1.05M | 63.4 |
Fable 5Anthropic · claude-fable-5 | $10 | $50 | $1 | 1M | 60.4 |
Grok 4.7xAI · grok-4.7 | $2† | $6† | $0.50 | 500K | 59.6 |
Qwen 3.8 MaxAlibaba · qwen3.8-max | $2 | $6 | $0.25 | 1M | 55.6 |
Claude Haiku 5.5Anthropic · claude-haiku-5-5 | $0.10† | $0.50† | $0.01 | 1M | 54.8 |
Grok 4.6xAI · grok-4.6 | $2† | $6† | $0.50 | 500K | 54.2 |
GLM-5.3Zhipu AI · glm-5.3 | $1.40 | $4.40 | $0.26 | 1M | 52.2 |
MiMo-V2.6 ProXiaomi · mimo-v2.6-pro | $0.435 | $0.87 | $0.0036 | 1.05M | 49.2 |
Muse Spark 1.3Meta · muse-spark-1.3 | $1.25 | $4.25 | $0.15 | 1.05M | 48.4 |
ChatGPT-5.6OpenAI · gpt-5.6 | $4† | $20† | $0.40 | 1.05M | 48.1 |
Gemini 3.8 FlashGoogle · gemini-3.8-flash | $0.75 | $3.75 | $0.075 | 1.05M | 43.8 |
GPT-5.5OpenAI · gpt-5.5 | $5† | $30† | $0.50 | 1.05M | 41.7 |
MiniMax M3.1 FlashMiniMax · MiniMax-M3.1-Flash | $0.15 | $0.60 | $0.015 | 1M | 38.5 |
Gemini 3.5 FlashGoogle · gemini-3.5-flash | $1.50 | $9 | $0.15 | 1.05M | 37.7 |
Grok 4.5xAI · grok-4.5 | $2† | $6† | $0.30 | 500K | 35 |
GPT-6 LunaOpenAI · gpt-6-luna | $0.10† | $0.50† | $0.01 | 1.05M | 32.8 |
GLM-5.2Zhipu AI · glm-5.2 | $1.40 | $4.40 | $0.26 | 1M | 31.4 |
Claude Opus 4.8Anthropic · claude-opus-4-8 | $5 | $25 | $0.50 | 1M | 30.4 |
Mistral Large 4Mistral · mistral-large-4 | $0.70 | $2.10 | $0.07 | 1.05M | 27.8 |
Kimi K3Moonshot · kimi-k3 | $3 | $15 | $0.30 | 1.05M | 27.1 |
GLM-5.3 FlashZhipu AI · glm-5.3-flash | $0.15 | $0.50 | $0.03 | 1M | 25 |
Claude Sonnet 5Anthropic · claude-sonnet-5 | $2 | $10 | $0.20 | 1M | 21.8 |
DeepSeek V4 ProDeepSeek · deepseek-v4-pro | $0.66 | $1.98 | $0.022 | 1.05M | 19.9 |
MiniMax M3MiniMax · MiniMax-M3 | $0.30† | $1.20† | $0.06 | 1M | 10.4 |
The published list prices in the models.dev catalog, refreshed daily; clouds and resellers may charge differently. The AW Index column is each model's standing on the bench index, which combines the public leaderboards.
More live models
| Model | Input | Output | Cached input | Context |
|---|---|---|---|---|
Muse Spark 1.1Meta · muse-spark-1.1 | $1.25 | $4.25 | $0.15 | 1.05M |
Muse Spark 1.2Meta · muse-spark-1.2 | $1.25 | $4.25 | $0.15 | 1.05M |
Gemini 3.6 FlashGoogle · gemini-3.6-flash | $0.75 | $3.75 | $0.075 | 1.05M |
Gemini 3.7 FlashGoogle · gemini-3.7-flash | $0.75 | $3.75 | $0.075 | 1.05M |
Gemini 3.5 Flash-LiteGoogle · gemini-3.5-flash-lite | $0.30 | $2.50 | $0.03 | 1.05M |
Qwen 3.8 FlashAlibaba · qwen3.8-flash | $0.15 | $0.47 | $0.016 | 1M |
MiMo-V2.5Xiaomi · mimo-v2.5 | $0.14 | $0.28 | $0.0028 | 1.05M |
† Long prompts cost more: GPT-6 Astra $20 / $75 over 272K tokens; GPT-6.1 Sol $4 / $15 over 272K tokens; GPT-6.1 Terra $2 / $7.50 over 272K tokens; GPT-6 Sol $4 / $15 over 272K tokens; Grok 4.7 $4 / $12 over 200K tokens; Claude Haiku 5.5 $0.50 / $2.50 over 100K tokens; Grok 4.6 $4 / $12 over 200K tokens; ChatGPT-5.6 $8 / $30 over 272K tokens; GPT-5.5 $10 / $45 over 272K tokens; Grok 4.5 $4 / $12 over 200K tokens; GPT-6 Luna $0.20 / $0.75 over 272K tokens; MiniMax M3 $0.60 / $2.40 over 512K tokens.
Price changes ahead
- Gemini 4 Argon goes from $2 / $10 to $4 / $20 per million tokens on December 1, 2026 (Google's pricing page)
- Gemini 3.6 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027 (Google's pricing page)
- Gemini 3.7 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027 (Google's pricing page)
- Gemini 3.8 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027 (Google's pricing page)
The tables above show today's rates.
Quick answers
Which LLM API is the cheapest?
Of the 40 live models priced here, the lowest input price is $0.10 per million tokens (Claude Haiku 5.5 and GPT-6 Luna) and the lowest output price $0.28 (MiMo-V2.5).
Which benchmarked model is the cheapest?
Among the 33 priced models scored on the ArtificialWatch bench index, the lowest output price is Claude Haiku 5.5, GPT-6 Luna and GLM-5.3 Flash share it, at $0.50 per million output tokens.
Which LLM API is the most expensive?
The highest output price of the 40 priced here is $50 per million tokens: GPT-6 Astra, Fable 5.1 and Fable 5 all charge it.
Where do these prices come from?
Each model maker's own API listing in the models.dev catalog, as of 2026-11-08, refreshed daily. A price is shown only for the catalog id that is the model's exact name, so a cheaper sibling tier never appears under a flagship's name. Clouds and resellers may charge differently; each model's page lists the providers that carry it.
Are any of these API prices about to change?
Yes, on the makers' own pricing pages: Gemini 4 Argon goes from $2 / $10 to $4 / $20 per million tokens on December 1, 2026; Gemini 3.6 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027; Gemini 3.7 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027; Gemini 3.8 Flash goes from $0.75 / $3.75 to $1.50 / $7.50 per million tokens on January 1, 2027. The table shows today's rates.
Price is one axis. For scores side by side, see the head-to-head comparisons and the bench index. For what shipped when, see new AI model releases. To hear when a new model goes live, watch free.
Use this chart
Writing about LLM API prices? Paste this where you want the chart. It redraws when the table updates, and the credit line links back to this page.
<figure><a href="https://artificialwatch.com/llm-pricing"><img src="https://artificialwatch.com/api/aw/og/llm-pricing" alt="What the top-ranked AI models cost per 1M tokens on their makers' own APIs, by ArtificialWatch" width="600" height="315"></a><figcaption>Chart: <a href="https://artificialwatch.com/llm-pricing">ArtificialWatch</a>, CC BY 4.0</figcaption></figure>Reports from this data
Nov 6, 2026
Google doubles seven Gemini API prices: Gemini 4 Argon reaches Claude Opus 5.5's $4 / $20 on December 1, and Gemini 3.8 Flash goes to $1.50 / $7.50 on January 1Nov 6, 2026
Prompt cache reads cost half as much on the newest AI models: 5% of the input price on Claude Opus 5.5, Sonnet 5.5, GPT-6.1 Sol and Gemini 4 Argon, against 10% a generation earlierNov 3, 2026
AI API prices don't follow the leaderboard: #1 Claude Opus 5.5 costs $4 / $20, while #3 and #4 cost $10 / $50
Every report, from every dataset: data reports.
Cite this page
Writing about LLM API prices? Copy a link or a citation.
<a href="https://artificialwatch.com/llm-pricing">LLM API pricing comparison</a>ArtificialWatch, "LLM API pricing comparison", https://artificialwatch.com/llm-pricingPrices move. Hear when a new model lands.
Free: a Chrome push and an email about 15 minutes after a new model answers on a public API. Members get a text or a phone call the moment a launch is confirmed.
- Free forever
- No card
- Unsubscribe in one click