Live data
Model prices
7 models, 1 of them free. Priced per million tokens so the numbers are comparable — providers quote per-token, which hides the difference.
| Model | Vendor | Input /1M | Output /1M | Context | Cutoff |
|---|---|---|---|---|---|
| Llama 4 Freefree | meta | $0.000 | $0.000 | 128K | 2026-03-01 |
| Gpt 5 Mini | openai | $0.150 | $0.600 | 200K | 2026-03-01 |
| Deepseek V4 | deepseek | $0.270 | $1.10 | 128K | 2026-03-01 |
| Gemini 3 Pro | $1.25 | $10 | 2M | 2026-03-01 | |
| Claude Sonnet 5 | anthropic | $3 | $15 | 400K | 2026-03-01 |
| Gpt 6 Astra | openai | $4 | $20 | 400K | 2026-03-01 |
| Claude Opus 5 | anthropic | $5 | $25 | 500K | 2026-03-01 |
Pricing comes from the OpenRouter API, which is a real interface rather than a scraped marketing page. A blank cell means the vendor publishes no price for that field — it does not mean zero.
Prices are per million tokens in US dollars, which is the only way to compare models fairly — providers quote per token, and the trailing zeros hide real differences.
About this data
What is the difference between input and output price?
Input is what you pay for the tokens you send (your prompt, your documents, the conversation so far). Output is what you pay for the tokens the model generates. Output is almost always the more expensive of the two, often by three to five times, so a chatty model can cost far more than its input price suggests.
What does the context window actually limit?
It is the maximum number of tokens the model can consider at once — your prompt and its reply combined. A 200K context roughly corresponds to a 500-page book. Bigger is not automatically better: you pay for every input token you send, so a large context is an option, not a target.
Why is a price cell blank?
Because the provider publishes no price for that field, not because it is free. Free models are labelled explicitly. We never render a missing price as zero.
What is the knowledge cutoff?
The date after which the model has no training knowledge of world events. It says nothing about whether the model can search the web — many can, which is how they answer questions about things after their cutoff.