GPT-4o (March 2025, chatgpt-4o-latest)

OpenAI

Verified against Artificial Analysis · Jul 13, 2026

Capability

percentile index

Price /M

$0 / $0

in / out

Context

—

Avg score

49.3

7 benchmarks

AA Intelligence

12.3

index

Weights

Closed

Released

Mar 27, 2025

About

GPT-4o (March 2025, chatgpt-4o-latest) is a proprietary model in the GPT family from OpenAI. Benchmarked on 7 evals, averaging 49.3.

Benchmark scores · 7

avg 49.3 — higher is better

MATHartificial-analysis · 2026-07

89.3

MMLU-Proartificial-analysis · 2026-07

80.3

GPQA Diamondartificial-analysis · 2026-07

65.5

LiveCodeBenchartificial-analysis · 2026-07

42.5

SciCodeartificial-analysis · 2026-07

36.6

AIMEartificial-analysis · 2026-07

25.7

Humanity's Last Examartificial-analysis · 2026-07

Details

PublisherOpenAI

FamilyGPT

Parameters—

License—

WeightsClosed

ReleasedMar 27, 2025

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$0/M