LLM API pricing, compared
Input and output prices per million tokens for 18 models, each read from the vendor’s own pricing page. Standard rates only: batch, cache and off-peak discounts are not applied.
Last checked October 22, 2026. Every vendor figure was read from that vendor’s own page on that date and links to it; a figure the page did not state is shown as “not stated”. Prices change, so confirm on the vendor’s page before you commit.
Price per million tokens
| Vendor | Model | Input | Output | Notes |
|---|---|---|---|---|
| Anthropic | Claude Fable 5.1 | $10.00 | $50.00 | |
| Anthropic | Claude Sonnet 5.5 | $2.00 | $10.00 | |
| Anthropic | Claude Haiku 5.5 | $0.10 | $0.50 | Input price for prompts up to 100,000 tokens. |
| DeepSeek | deepseek-v4-pro | $1.32 | $3.96 | Peak-hour rate, cache miss. Off-peak rates are half. |
| DeepSeek | deepseek-flash | $0.30 | $1.20 | Peak-hour rate, cache miss. Off-peak rates are half. |
| Gemini 4 Argon | $2.00 | $10.00 | Listed through November 30, 2026; $4.00 in and $20.00 out from December 1, 2026. Not on the free tier. | |
| Gemini 3.8 Flash | $0.75 | $3.75 | Listed through December 31, 2026; $1.50 in and $7.50 out from January 1, 2027. | |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 | Text, image and video input; audio input is $0.50. | |
| Groq | openai/gpt-oss-120b | $0.15 | $0.60 | |
| Groq | openai/gpt-oss-20b | $0.075 | $0.30 | |
| Mistral | Mistral Large 4 | $1.36 | $4.18 | |
| Mistral | Mistral Medium 3.5 | $1.50 | $7.50 | |
| Mistral | Ministral 3 3B | $0.10 | $0.10 | |
| OpenAI | gpt-6-astra | $10.00 | $50.00 | Short-context tier (up to 272K input tokens). |
| OpenAI | gpt-6.1-sol | $2.00 | $10.00 | Short-context tier (up to 272K input tokens). |
| OpenAI | gpt-6-luna | $0.10 | $0.50 | Short-context tier (up to 272K input tokens). |
| xAI | grok-4.7 | $2.00 | $6.00 | Under 200K tokens of context; $4.00 in and $12.00 out above. |
| xAI | grok-4.3 | $1.25 | $2.50 |
- Anthropic free tier: New users receive a small amount of free credits to test the API. Source
- Google free tier: A free tier with free input and output tokens on some models (Gemini 3.8 Flash and 3.1 Flash-Lite, not Gemini 4 Argon). Source
- Groq free tier: A free plan with lower rate limits (30 requests per minute on the models above). Source
- No free tier stated: DeepSeek, Mistral, OpenAI, xAI, on the pricing pages read.
Questions
How is LLM API pricing quoted?
Per million tokens, with separate prices for input (the prompt) and output (the reply). This table lists each vendor's standard rate: batch processing, prompt caching and off-peak discounts lower it, and some vendors charge more above a context-length threshold, which the notes column names.
What is the price range across these models?
As listed on October 22, 2026, input prices in this table start at $0.075 per million tokens (openai/gpt-oss-20b) and output prices go up to $50.00 (Claude Fable 5.1). A lower list price says nothing about quality or speed, so compare on your own workload.
Are any of these prices about to change?
Yes. Google lists Gemini 4 Argon at this price through November 30, 2026; Google lists Gemini 3.8 Flash at this price through December 31, 2026. The notes column gives the price after that date.
Does upAPI sell LLM API access?
No. upAPI's public catalog is data and automation APIs (search, scraping, maps, social, developer tools); it does not resell model tokens. This table is a reference for the agents and apps that call both.