LLM API prices by provider
Price ranges per 1M tokens across each provider's current models. Jump to a provider for every model.
| Provider | Models | Input per 1M | Output per 1M | Batch | Free tier |
|---|---|---|---|---|---|
| OpenAI | 24 | $0.05 to $30.00 | $0.40 to $180.00 | 50% off | Is OpenAI API free? |
| Anthropic | 14 | $0.10 to $10.00 | $0.50 to $50.00 | 50% off | Is Claude API free? |
| 9 | $0.25 to $2.00 | $1.50 to $12.00 | 50% off | Is Google Gemini API free? | |
| xAI | 3 | $1.25 to $2.00 | $2.50 to $6.00 | Not verified | Not tracked |
| DeepSeek | 2 | $0.30 to $1.32 | $1.20 to $3.96 | Not verified | Is DeepSeek API free? |
| Mistral | 3 | $0.15 to $0.70 | $0.60 to $2.10 | Not verified | Is Mistral free? |
| Meta Llama | 2 | $0.10 to $0.1875 | $0.32 to $0.6525 | Not verified | Not tracked |
Need it for $0 while you prototype? 5 providers still have a standing free tier; see the free LLM APIs we track, with their real limits.
OpenAI API prices
| Model | Input | Cached input | Output | Batch input | Batch output |
|---|---|---|---|---|---|
| GPT-6 Astra | $10.00 | $1.00 | $50.00 | $5.00 | $25.00 |
| GPT-6.1 Sol | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 |
| GPT-6.1 Terra | $1.00 | $0.05 | $5.00 | $0.50 | $2.50 |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 | $0.05 | $0.25 |
| GPT-5.6 Solpromo to 30 Nov | $4.00 | $0.40 | $20.00 | $2.00 | $10.00 |
| GPT-5.6 Terra | $2.00 | $0.20 | $12.00 | $1.00 | $6.00 |
| GPT-5.6 Luna | $0.20 | $0.02 | $1.20 | $0.10 | $0.60 |
| GPT-5.5 Pro | $30.00 | Not listed | $180.00 | $15.00 | $90.00 |
| GPT-5.5 | $5.00 | $0.50 | $30.00 | $2.50 | $15.00 |
| GPT-5.4 Pro | $30.00 | Not listed | $180.00 | $15.00 | $90.00 |
| GPT-5.4 | $2.50 | $0.25 | $15.00 | $1.25 | $7.50 |
| GPT-5.4 Mini | $0.75 | $0.075 | $4.50 | $0.375 | $2.25 |
| GPT-5.2 Pro | $21.00 | Not listed | $168.00 | $10.50 | $84.00 |
| GPT-5.2 | $1.75 | $0.175 | $14.00 | $0.875 | $7.00 |
| GPT-5.1retires 1 Apr 2027 | $1.25 | $0.125 | $10.00 | $0.625 | $5.00 |
| GPT-5retires 11 Dec 2026 | $1.25 | $0.125 | $10.00 | $0.625 | $5.00 |
| GPT-5 Miniretires 11 Dec 2026 | $0.25 | $0.025 | $2.00 | $0.125 | $1.00 |
| GPT-5 Nanoretires 11 Dec 2026 | $0.05 | $0.005 | $0.40 | $0.025 | $0.20 |
| o3-proretires 11 Dec 2026 | $20.00 | Not listed | $80.00 | $10.00 | $40.00 |
| o3retires 11 Dec 2026 | $2.00 | $0.50 | $8.00 | $1.00 | $4.00 |
| GPT-4.1retires 14 Jan 2027 | $2.00 | $0.50 | $8.00 | $1.00 | $4.00 |
| GPT-4.1 Miniretires 14 Jan 2027 | $0.40 | $0.10 | $1.60 | $0.20 | $0.80 |
| GPT-4oretires 14 Jan 2027 | $2.50 | $1.25 | $10.00 | $1.25 | $5.00 |
| GPT-4o Miniretires 14 Jan 2027 | $0.15 | $0.075 | $0.60 | $0.075 | $0.30 |
Per 1M tokens, from OpenAI's pricing page, checked 6 Nov 2026. GPT-5.4 and later charge a higher rate above 272K input tokens. The Pro models have no cached price. A cost ranking of every OpenAI model, and the extra charges to plan for, are on the OpenAI API pricing page. To size a prompt before you pay, count OpenAI tokens.
Claude API prices (Anthropic)
| Model | Input | Cached input | Output | Batch input | Batch output |
|---|---|---|---|---|---|
| Claude Fable 5.1 | $10.00 | $0.25 | $50.00 | $5.00 | $25.00 |
| Claude Opus 5.5 | $4.00 | $0.20 | $20.00 | $2.00 | $10.00 |
| Claude Sonnet 5.5 | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 |
| Claude Haiku 5.5 | $0.10 | $0.01 | $0.50 | $0.05 | $0.25 |
| Claude Fable 5 | $10.00 | $1.00 | $50.00 | $5.00 | $25.00 |
| Claude Opus 5 | $5.00 | $0.50 | $25.00 | $2.50 | $12.50 |
| Claude Sonnet 5 | $2.00 | $0.20 | $10.00 | $1.00 | $5.00 |
| Claude Opus 4.8 | $5.00 | $0.50 | $25.00 | $2.50 | $12.50 |
| Claude Opus 4.7 | $5.00 | $0.50 | $25.00 | $2.50 | $12.50 |
| Claude Opus 4.6 | $5.00 | $0.50 | $25.00 | $2.50 | $12.50 |
| Claude Sonnet 4.6 | $3.00 | $0.30 | $15.00 | $1.50 | $7.50 |
| Claude Opus 4.5 | $5.00 | $0.50 | $25.00 | $2.50 | $12.50 |
| Claude Sonnet 4.5retires 30 Nov 2026 | $3.00 | $0.30 | $15.00 | $1.50 | $7.50 |
| Claude Haiku 4.5 | $1.00 | $0.10 | $5.00 | $0.50 | $2.50 |
Per 1M tokens, from Anthropic's pricing page, checked 6 Nov 2026. Claude Haiku 5.5 charges a higher rate above 100K input tokens. Claude Opus 4.7 and later use a tokenizer that turns the same text into about 30% more tokens than earlier Claude models, so a lower price per token does not always mean a lower bill. A cost ranking of every Anthropic model, and the extra charges to plan for, are on the Claude API pricing page. To size a prompt before you pay, count Claude tokens.
Gemini API prices (Google)
| Model | Input | Cached input | Output | Batch input | Batch output |
|---|---|---|---|---|---|
| Gemini 4 Argonpromo to 30 Nov | $2.00 | $0.10 | $10.00 | $1.00 | $5.00 |
| Gemini 3.8 Flashpromo to 31 Dec | $0.75 | $0.075 | $3.75 | $0.375 | $1.875 |
| Gemini 3.6 Flashpromo to 31 Dec | $0.75 | $0.075 | $3.75 | $0.375 | $1.875 |
| Gemini 3.5 Flash Lite | $0.30 | $0.03 | $2.50 | $0.15 | $1.25 |
| Gemini 3.1 Flash Literetires 7 May 2027 | $0.25 | $0.025 | $1.50 | $0.125 | $0.75 |
| Gemini 3.1 Pro | $2.00 | $0.20 | $12.00 | $1.00 | $6.00 |
| Gemini 3 Flash | $0.50 | $0.05 | $3.00 | $0.25 | $1.50 |
| Gemini 2.5 Proretires 1 Dec 2026 | $1.25 | $0.125 | $10.00 | $0.625 | $5.00 |
| Gemini 2.5 Flashretires 1 Dec 2026 | $0.30 | $0.03 | $2.50 | $0.15 | $1.25 |
Per 1M tokens, from Google's Gemini API pricing page, checked 6 Nov 2026. Prices are the paid tier of the Gemini Developer API. Thinking tokens are billed as output. A cost ranking of every Google model, and the extra charges to plan for, are on the Gemini API pricing page. To size a prompt before you pay, count Gemini tokens.
Grok API prices (xAI)
| Model | Input | Cached input | Output |
|---|---|---|---|
| Grok 4.7 | $2.00 | $0.50 | $6.00 |
| Grok 4.5 | $2.00 | $0.30 | $6.00 |
| Grok 4.20 | $1.25 | $0.20 | $2.50 |
Per 1M tokens, from xAI's models and pricing page, checked 6 Nov 2026. Prices are for prompts under 200K tokens; longer prompts cost more. We have not verified a batch discount for this provider.
DeepSeek API prices
| Model | Input | Cached input | Output |
|---|---|---|---|
| DeepSeek V4.1 Flash | $0.30 | $0.006 | $1.20 |
| DeepSeek V4 Pro | $1.32 | $0.044 | $3.96 |
Per 1M tokens, from DeepSeek's pricing page, checked 6 Nov 2026. These are peak-hour prices, which apply 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays. Every other hour, including weekends and Chinese public holidays, costs half. We have not verified a batch discount for this provider.
Mistral API prices
| Model | Input | Cached input | Output |
|---|---|---|---|
| Mistral Large 4 | $0.70 | $0.07 | $2.10 |
| Mistral Small 4 | $0.15 | $0.015 | $0.60 |
| Mistral Large 3retires 30 Apr 2027 | $0.50 | Not listed | $1.50 |
Per 1M tokens, from Mistral's API pricing page, checked 6 Nov 2026. Mistral Large 4 left preview on 6 October 2026. Mistral Large 3 is deprecated. We have not verified a batch discount for this provider.
Llama API prices (Meta, via OpenRouter)
| Model | Input | Cached input | Output |
|---|---|---|---|
| Llama 4 Maverick | $0.1875 | Not listed | $0.6525 |
| Llama 3.3 70B | $0.10 | Not listed | $0.32 |
Per 1M tokens, from OpenRouter's Llama listings, checked 6 Nov 2026. Meta does not sell API access to Llama directly. These are the rates OpenRouter lists. We have not verified a batch discount for this provider.
What does a typical workload cost?
Take a small chatbot: 1,000 input and 500 output tokens a request, 100 requests a day, for 30 days. That is 3M input and 1.5M output tokens a month. Here is the monthly bill at list prices on the five cheapest models and on the four flagships from our AI model comparison, which covers context, speed and reasoning.
| Model | Provider | Per request | Per month |
|---|---|---|---|
| GPT-5 Nanoretires 11 Dec 2026 | OpenAI | $0.00025 | $0.75 |
| Llama 3.3 70B | Meta | $0.00026 | $0.78 |
| GPT-6 Luna | OpenAI | $0.00035 | $1.05 |
| Claude Haiku 5.5 | Anthropic | $0.00035 | $1.05 |
| GPT-4o Miniretires 14 Jan 2027 | OpenAI | $0.00045 | $1.35 |
| GPT-6.1 Sol | OpenAI | $0.007 | $21.00 |
| Claude Opus 5.5 | Anthropic | $0.014 | $42.00 |
| Gemini 4 Argonpromo to 30 Nov | $0.007 | $21.00 | |
| Grok 4.7 | xAI | $0.005 | $15.00 |
The spread is the point: the same traffic costs $0.75 or $42.00 a month depending on the model. Change any of these numbers, or turn on caching and batch, in the LLM cost calculator.
How does LLM API pricing work?
You pay per token, at one price for the tokens you send (input) and a higher one for the tokens the model writes (output). Five things change the bill beyond the list price:
- Output costs more. On the models here, output is 5 times the input price in the middle of the range and up to 8.3 times at the top. On reasoning models, the thinking tokens count as output too.
- Cached input is cheaper. When the start of a prompt repeats, such as a long system prompt, providers bill those tokens at a cached rate. It is 10% of the input price on 34 of the 50 models that list one. For a measured example, see what caching saved on a real workload.
- Batch is half price. OpenAI, Anthropic and Google charge 50% less for input and output sent through their batch APIs, which return results later instead of right away.
- Promotions end. 4 models are on temporary prices right now. We mark them "promo" in the tables and list each end date, with the price that follows, under "Promo prices ending soon".
- Tokens are not the same size everywhere. The same text becomes a different number of tokens on each model, so compare on your own prompt with the token counter.
How often do LLM API prices change?
Often. Our pricing changelog has 5 updates with price changes in 2026, the latest on 6 Nov 2026. Launch promotions are the most common change: a new model ships at a discount, and the list price starts on a fixed date. We re-check every price against the provider's own page; here is how we verify prices.
Frequently asked
How much does 1M tokens cost?
It depends on the model and on whether the tokens are input or output. Across the 57 models we track, input costs from $0.05 per 1M tokens (GPT-5 Nano) to $30.00 (GPT-5.4 Pro), and output from $0.32 (Llama 3.3 70B) to $180.00 (GPT-5.4 Pro). Models we class as flagships charge $0.50 to $5.00 per 1M input tokens. Checked 6 Nov 2026.
Why do output tokens cost more than input tokens?
The model reads your whole prompt in one parallel pass, but it writes the reply one token at a time, so each output token takes more compute. On the models here, output costs 5 times the input price in the middle of the range, and anywhere from 2 to 8.3 times. Capping reply length is one of the quickest ways to cut your API bill.
How do I work out my monthly LLM API cost?
Multiply input tokens by the input price and output tokens by the output price, divide by 1,000,000, then multiply by your request count. For example, 1,000 input and 500 output tokens on GPT-6.1 Sol cost $0.007 a request, or $21.00 a month at 100 requests a day. The LLM cost calculator does this for every model at once, with caching and batch.
Are batch API prices cheaper?
Yes, at OpenAI, Anthropic and Google: their batch APIs charge half the normal input and output price for requests that can wait for results. We have not verified a batch discount for xAI, DeepSeek, Mistral or the Llama hosts, so the tables show batch prices only for the first three.