Compare

o4 vs Claude Sonnet 4.6

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

Verdict

Claude Sonnet 4.6 is ~4.4× cheaper on a 3:1 input:output blend.

OpenAI

Deep-reasoning model. Spends more tokens thinking to crush math, science, and hard coding problems.

State-of-the-art math and scienceExtremely high-ceiling reasoning

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$15.00$3.00

Output / 1M$60.00$15.00

Context200K500K

Max output100K64K

LicenseProprietaryProprietary

Released2026-02-182025-11-04

Benchmarks

LMSYS ELO—1368.0

MMLU Pro93.888.7

HumanEval92.392.1

SWE-Bench68.064.5

MATH96.285.0

GPQA82.559.4

MMMU—74.1

IFEval—90.5

Per-task cost

Summarize a 1-hour meeting transcript$0.255$0.052

Review a 500-line pull request$0.240$0.054

Answer a customer support ticket$0.093$0.021

Extract structured data from a resume$0.135$0.030

Debug a stack trace with context$0.210$0.048

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

Add up to four models, tweak tasks live.