Which is better, Claude Sonnet 4-20250514 (no thinking) or GPT-4o?

Refresh Models —Deals —Regulatory —Sources —

Questions this page answers

Which is better, Claude Sonnet 4-20250514 (no thinking) or GPT-4o?: Across 6 shared benchmarks, Claude Sonnet 4-20250514 (no thinking) scores higher on 4 and GPT-4o on 2. The widest gap is aider_polyglot, where Claude Sonnet 4-20250514 (no thinking) scores 56.4 against 30.7.

Claude Sonnet 4-20250514 (no thinking) vs GPT-4o

Across 6 shared benchmarks, Claude Sonnet 4-20250514 (no thinking) scores higher on 4 and GPT-4o on 2. The widest gap is aider_polyglot, where Claude Sonnet 4-20250514 (no thinking) scores 56.4 against 30.7.

AnthropicvsOpenAI6 shared benchmarks4–2 head-to-head

Benchmark	Claude Sonnet 4-20250514 (no thinking)	GPT-4o
aider_polyglot	56.4	30.7
arena_vision	1176	1162
vectara_answer_rate	98.6	93.8
vectara_avg_summary_length	145.8	86.6
vectara_factual_consistency	89.7	90.4
vectara_hallucination_rate ↓	10.3	9.6

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.