Compare official and relay input/output token prices across OpenAI, Anthropic, Google, DeepSeek and more — with source and update status.
662
models
65
providers
$0.00055
lowest input/M
-99%
max relay savings
3697 live-scraped (); 288 manually maintained
Example: 1,000 requests, each with 1,000 input and 200 output tokens. That is 1M input + 0.2M output tokens. These are uncached token subtotals, not final bills.
Formula: input tokens / 1M × input rate + output tokens / 1M × output rate. Cache writes, long-context tiers, tool fees, taxes and per-request minimums are excluded. An unknown minimum is not zero. Check quote dates and provider terms, then validate against a real bill. USD is the comparison currency; CNY quotes use the site's fixed reference conversion.
Compare token costs for your own usage →Best prices for popular models
Same-tier model comparisons
Ordered by the Artificial Analysis 2026-09-09 snapshot; 10 featured vendors appear first and the rest stay folded.
External Elo is advisory; scores from different task leaderboards are not directly comparable. Each detail page identifies its board and date. Verify price source and update time there. View leaderboard source
Ordered by the Artificial Analysis 2026-08-17 snapshot; 10 featured vendors appear first and the rest stay folded.
External Elo is advisory; scores from different task leaderboards are not directly comparable. Each detail page identifies its board and date. Verify price source and update time there. View leaderboard source
Ordered by the Artificial Analysis 2026-08-17 snapshot; 5 featured vendors appear first and the rest stay folded.
External Elo is advisory; scores from different task leaderboards are not directly comparable. Each detail page identifies its board and date. Verify price source and update time there. View leaderboard source
Dedicated category pages
Four categories: LLMs (language models), video generation, image generation, and audio & speech (TTS/ASR). Use the tabs above to browse each category, or visit the dedicated hub pages for per-model detail comparisons.
For relay platforms with live API access (such as OpenRouter), LLM prices auto-scrape every ~6 hours. Official providers (OpenAI, Anthropic, etc.) and the Video/Audio/Image categories are manually maintained from official pricing pages and updated when providers announce changes.
Official providers are the model developers themselves (OpenAI, Anthropic, Google, etc.). Relay services resell API access — often at a discount, but with different latency, SLA, and support terms. ComputeUnion lists both so you can choose based on your priorities.
No provider is cheapest for every model. The lowest current input quote is $0.00055/1M tokens; compare output cost, latency, and SLA before buying.