LegalBench ↗82.21 %
GPT 4o (2024-11-20) · LegalBench · Accuracy ↗
OpenAI / MODEL EVIDENCE
GPT-4o (2024-11-20)
16 benchmarks · 18 observations · 16 sources
Scores at a glance
16 benchmarks · large dot: GPT-4o (2024-11-20); small dots: closest models where reported
CaseLaw v2 ↗59.7 %
GPT 4o (2024-11-20) · CaseLaw · Accuracy ↗
CorpFin v2 ↗45.92 %
GPT 4o (2024-11-20) · CorpFin · Accuracy ↗
MortgageTax & TaxEval v2 ↗57.43 %
GPT 4o (2024-11-20) · MortgageTax · Accuracy ↗
1 more reported result
74.53 % · GPT 4o (2024-11-20) · TaxEval · Accuracy · Vals AI · TaxEval ↗
AIME 2024–2025 · Vals ↗11.88 %
GPT 4o (2024-11-20) · AIME · Accuracy ↗
MATH-500 ↗74 %
GPT 4o (2024-11-20) · MATH 500 · Accuracy ↗
MGSM (Multilingual GSM) ↗90.36 %
GPT 4o (2024-11-20) · MGSM · Accuracy ↗
GPQA Diamond ↗53.79 %
GPT 4o (2024-11-20) · GPQA Diamond · Accuracy ↗
MMLU-Pro ↗72.56 %
GPT 4o (2024-11-20) · MMLU Pro · Accuracy ↗
MMMU-Pro ↗62.16 %
GPT 4o (2024-11-20) · MMMU Pro · Accuracy ↗
LiveCodeBench ↗43.44 %
GPT 4o (2024-11-20) · LiveCodeBench · Accuracy ↗
gpt-4o-2024-11-20 · Aider coding benchmarks · Polyglot · percent correct ↗
1 more reported result
71.4% · gpt-4o-2024-11-20 · Aider coding benchmarks · Refactoring · percent correct · Aider refactoring benchmark ↗
SimpleQA ↗38.8%
gpt-4o-2024-11-20 · SimpleQA · Correct answers · OpenAI Simple Evals · archived July 2025 ↗
HumanEval ↗90.2%
gpt-4o-2024-11-20 · HumanEval · Pass@1 · OpenAI Simple Evals · archived July 2025 ↗
gpt-4o-2024-11-20 · DROP (Discrete Reasoning Over Paragraphs) · F1 · three-shot · OpenAI Simple Evals · archived July 2025 ↗
gpt-4o-2024-11-20 · FrontierMath-2025-02-28-Public · FrontierMath public · 10 questions · source run score · Epoch AI · public February 2025 sample · historical release ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
GPT-4o (2024-08-06) →OpenAI
8 of 11 shared benchmark comparisons
7 of 11 shared benchmark comparisons
GPT-4o Mini →OpenAI
9 of 11 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- Aider code editing benchmark ↗
- Aider refactoring benchmark ↗
- A Rosetta Stone for AI Benchmarks — released Epoch evaluation runs ↗
- Simple Evals — model benchmark results ↗
- AIME — evaluation methodology and results ↗
- CaseLaw — evaluation methodology and results ↗
- CorpFin — evaluation methodology and results ↗
- GPQA Diamond — evaluation methodology and results ↗
- LiveCodeBench — evaluation methodology and results ↗
- LegalBench — evaluation methodology and results ↗
- MATH 500 — evaluation methodology and results ↗
- MGSM — evaluation methodology and results ↗
- MMLU Pro — evaluation methodology and results ↗
- MMMU Pro — evaluation methodology and results ↗
- MortgageTax — evaluation methodology and results ↗
- TaxEval — evaluation methodology and results ↗
Names used in source tables
GPT 4o (2024-11-20) · gpt-4o-2024-11-20