Compare

GPT-5 Mini vs Claude Sonnet 4.6

OpenAI vs Anthropic. Specs, benchmarks, and real per-task cost — all in one page.

Verdict

Claude Sonnet 4.6 leads on LMSYS ELO (1368 vs 1288). GPT-5 Mini is ~6.9× cheaper on a 3:1 input:output blend.

OpenAI

GPT-5’s smaller, cheaper sibling. Optimized for chat and lightweight agent tasks.

Very low latencyCheap enough for mass consumer apps

Anthropic

The workhorse. Near-Opus quality at 1/5 the cost. The default choice for production code and agent workloads.

Excellent cost/quality ratioFast inferenceGreat for Cursor/Claude Code

Pricing

Input / 1M$0.50$3.00

Output / 1M$2.00$15.00

Context256K500K

Max output16K64K

LicenseProprietaryProprietary

Released2026-01-212025-11-04

Benchmarks

LMSYS ELO1288.01368.0

MMLU Pro78.588.7

HumanEval85.592.1

SWE-Bench—64.5

MATH79.885.0

GPQA—59.4

MMMU—74.1

IFEval85.490.5

Per-task cost

Summarize a 1-hour meeting transcript$0.0085$0.052

Review a 500-line pull request$0.0080$0.054

Answer a customer support ticket$0.0031$0.021

Extract structured data from a resume$0.0045$0.030

Debug a stack trace with context$0.0070$0.048

Per-call cost using published token counts for each task. Real-world prompts vary.

Add up to four models, tweak tasks live.