StackProofStackProof

LLM price comparisons: every model head-to-head (August 2026)

Pick any two models. Each page computes the exact per-task gap from the open dataset, so the comparison stays honest when prices move.

Mistral Small ($0.00158/task)

DeepSeek-V4-Flash ($0.00396/task)

Gemini 2.5 Flash ($0.00510/task)

Grok 4.3 ($0.00938/task)

GPT-5.4 mini ($0.01013/task)

Claude Haiku 4.5 ($0.01200/task)

Claude Sonnet 5 ($0.02400/task)

Gemini 3.1 Pro ($0.02700/task)

Claude Opus 4.8 ($0.06000/task)