GPT-4 (Jun 2023)

Capability & preference

General capability · ECI

A cross-benchmark capability estimate. Index points, not a percentage; intervals show uncertainty.

GPT-4 (Jun 2023) 120.2^* GPT-4 (Jun 2023) · 95% interval 116.6–123.5

Model performance

Radar task score: four tests, equal weight. An experimental summary of these tasks, not a universal intelligence rating.

Tested configuration Radar task score

Unrated

2/4 core tests · not enough evidence

3 benchmark results for this configuration

Scientific reasoning

GPQA Diamond · 25%

Expert-written science questions; strict multiple-choice scoring. Chance accuracy is 25%.

GPT-4 (Jun 2023) 30.7^*% Approx. 95% interval: 26.1–35.2% · 2025-01-27 · Older result

Novel problem solving

Mystery Game Puzzles · 25%

Best-next-move puzzles in an undisclosed game variant. A narrow reasoning test, not general intelligence.

GPT-4 (Jun 2023) 12.0^*% Approx. 95% interval: 5.6–18.4% · 2026-08-28

Advanced mathematics

FrontierMath Tiers 1-3 · 25%

Private v2 mathematics problems with Python tools. Earlier versions and Tier 4 are excluded.

GPT-4 (Jun 2023) Not available

Factual answers

SimpleQA Verified · 25%

Correct factual answers under a guessing prompt, graded by a model. Not a hallucination or abstention score.

GPT-4 (Jun 2023) Not available

Competition mathematics

OTIS Mock AIME 2024-2025 · Separate result

Competition problems with model-extracted answers checked against an exact key. Earlier grading procedures are excluded.

GPT-4 (Jun 2023) 1.1^*% Approx. 95% interval: 0.1–2.2% · 2025-10-23 · Older result

Bars show reported task scores. Whiskers are approximate 95% intervals where standard errors are available. Different reasoning budgets are not equal-cost tests. Missing results are never zero.

Run it locally

Inference · memory estimate

GPT-4 (Jun 2023)

Closed weights. No supported local deployment or GPU requirement is published.

Capabilities

Open weights: No^*

System configuration

Detailed specifications

Specifications · 4 fields

Access model: API access^*
Open weights: No^*
Organization: OpenAI^*
Organization country: United States of America^*