LLM API Prices: Claude, GPT and Gemini
Prices per 1 million tokens for 12 Claude, GPT and Gemini models, checked on October 14, 2026 against each provider's official pricing page. The lowest input price here is $0.10 (Claude Haiku 5.5 and GPT-6 Luna). Batch requests cost half the standard price at all three providers, and cached input costs a fraction of normal input. To turn these prices into a monthly bill, use the LLM API cost calculator.
Download as CSV| Model | Input | Cached input | Output | Batch input | Batch output | Notes |
|---|---|---|---|---|---|---|
| Claude Fable 5.1Anthropic | $10.00 | $0.25 | $50.00 | $5.00 | $25.00 | |
| Claude Opus 5.5Anthropic | $4.00 | $0.20 | $20.00 | $2.00 | $10.00 | |
| Claude Sonnet 5.5Anthropic | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 | |
| Claude Haiku 5.5Anthropic | $0.10 | $0.01 | $0.50 | $0.05 | $0.25 | Prompts over 100k tokens: $0.50 input, $0.05 cached, $2.50 output |
| Claude Haiku 4.5Anthropic | $1.00 | $0.10 | $5.00 | $0.50 | $2.50 | |
| GPT-6 AstraOpenAI | $10.00 | $1.00 | $50.00 | $5.00 | $25.00 | |
| GPT-6.1 SolOpenAI | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 | |
| GPT-6 LunaOpenAI | $0.10 | $0.01 | $0.50 | $0.05 | $0.25 | |
| Gemini 4 ArgonGoogle | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 | Introductory price until Nov 30, 2026 ($4 / $20 after) |
| Gemini 3.1 Pro PreviewGoogle | $2.00 | $0.20 | $12.00 | $1.00 | $6.00 | Prompts over 200k tokens: $4.00 input, $0.40 cached, $18.00 output |
| Gemini 3.8 FlashGoogle | $0.75 | $0.075 | $3.75 | $0.375 | $1.875 | Introductory price until Dec 31, 2026 ($1.50 / $7.50 after) |
| Gemini 3.5 Flash-LiteGoogle | $0.30 | $0.03 | $2.50 | $0.15 | $1.25 |
Standard list prices in USD from the official pages of Anthropic, OpenAI and Google. Cache write premiums, storage fees, regional surcharges and taxes are not included. Always confirm on the provider's page before committing to a budget.
How to read these prices
- Input and output are charged separately. Output usually costs several times more, so long answers drive the bill.
- Cached input applies to the repeated part of a prompt that the provider has stored, such as long instructions or a document you ask about many times.
- Batch prices apply to requests sent through each provider's Batch API, which returns results within 24 hours instead of right away.
- Long-prompt tiers raise the price of the whole request once a single prompt passes the stated size.
Change log
All prices re-checked on the official pages. Added Claude Haiku 5.5 (with its higher price above 100k-token prompts) and Gemini 4 Argon (introductory price until November 30, 2026). Added cached-input and batch prices for every model.
First published with 10 models from Anthropic, OpenAI and Google.
Building an AI feature?
Prices are only half of the cost: prompt size, caching and the right model per task matter as much. I build LLM features and workflows that stay within budget; see AI workflow automation or hire me on Upwork.