Compare

Claude Sonnet 4.6 vs GPT-5 Mini

Anthropic vs OpenAI. 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.

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

OpenAI

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

Very low latencyCheap enough for mass consumer apps

Pricing

Input / 1M$3.00$0.50

Output / 1M$15.00$2.00

Context500K256K

Max output64K16K

LicenseProprietaryProprietary

Released2025-11-042026-01-21

Benchmarks

LMSYS ELO1368.01288.0

MMLU Pro88.778.5

HumanEval92.185.5

SWE-Bench64.5—

MATH85.079.8

GPQA59.4—

MMMU74.1—

IFEval90.585.4

Per-task cost

Summarize a 1-hour meeting transcript$0.052$0.0085

Review a 500-line pull request$0.054$0.0080

Answer a customer support ticket$0.021$0.0031

Extract structured data from a resume$0.030$0.0045

Debug a stack trace with context$0.048$0.0070

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

Add up to four models, tweak tasks live.