GPT-4o (source unspecified)

Name: GPT-4o (source unspecified)
Author: OpenAI

OpenAI

This benchmark entry retains GPT-4o observations whose source names an unpinned GPT-4o route or omits the evaluated snapshot. OpenAI released distinct May 13, August 6, and November 20, 2024 snapshots; results that identify one of those artifacts are attached to its separate model row, while this entry does not assert a shared checkpoint or release date.

Benchmark scores

Benchmark	Score
A-Fantasia	48.0
A-Fantasia (Valid Attempts with Retries)	47.9
A-Fantasia Backwards Spelling	57.0
A-Fantasia Chess	13.0
A-Fantasia Chess (Valid Attempts with Retries)	13.0
A-Fantasia Cube (Valid Attempts with Retries)	74.0
A-Fantasia Cube Rotation	74.0
A-Fantasia Spelling (Valid Attempts with Retries)	56.6
Active Probing Benchmark (APB)	19.6
Active Probing Benchmark (Biology)	26.7
Active Probing Benchmark (Chemistry)	20.0
Active Probing Benchmark (Computer Science)	10.8
Active Probing Benchmark (Economics)	25.0
Active Probing Benchmark (Human Disease)	24.2
Active Probing Benchmark (Mathematics)	11.7
Active Probing Benchmark (Philosophy)	20.8
Active Probing Benchmark (Physics)	17.5
AgentVQA	31.4
AgentVQA Egocentric Videos	40.0
AgentVQA Games	29.0
AgentVQA Robotics	31.5
AgentVQA Spatial Planning	35.0
AgentVQA Web Agents	21.5
AIME 2025 (MathArena repeated-run mean)	11.7
Blueprint-Bench	0.1
BuseyBench SVG	0.3
CadQueryEval	24.0
CadQueryEval — Bounding Box	44.0
CadQueryEval — Chamfer Distance	48.0
CadQueryEval — Code Execution	60.0