LLM Pricing 2026 — Updated every 6 hours from official sources

Last updated: . Data licensed for agent consumption via x402 ($0.001/call).

Try it as an agent:

curl https://devrefs.dev/api/llm-prices?model=claude-opus-4-7

12 modèles tracked

Provider Model Input $/MTok Output $/MTok Effective Cost Factor Date Modified Source
Anthropicclaude-opus-4-715.0075.001.352026-05-05anthropic.com
Anthropicclaude-sonnet-4-63.0015.001.002026-05-05anthropic.com
Anthropicclaude-haiku-4-50.251.251.002026-05-05anthropic.com
OpenAIgpt-55.0015.001.002026-05-05openai.com
OpenAIgpt-4o2.5010.001.002026-05-05openai.com
OpenAIgpt-4o-mini0.150.601.002026-05-05openai.com
Googlegemini-2-5-pro1.255.001.002026-05-05ai.google.dev
Googlegemini-2-0-flash0.0750.301.002026-05-05ai.google.dev
Mistralmistral-large-32.006.001.002026-05-05mistral.ai
Mistralmistral-small-30.200.601.002026-05-05mistral.ai
DeepSeekdeepseek-r20.552.191.002026-05-05deepseek.com
DeepSeekdeepseek-v30.271.101.002026-05-05deepseek.com

Note : table fallback values displayed when prerendered. The live /api/llm-prices endpoint always returns cron-fresh values from KV (≤ 6h freshness, with fetched_at + HMAC _signature).

For agents

Atomic JSON payloads, 1 query = 1 model, ≤ 50 KB. dateModified + HTTP Last-Modified + fetched_at = triple freshness signal.

Back to home OpenAPI 3.1