GPT-4.5 — benchmark results

OpenAI's large GPT-4.5 non-reasoning model (February 2025), later retired from the API. Provider: OpenAI. Released 2025-02-27. Access: API.

Unified ELO 1586 ± 8, rank #510 of 1776 rated models, from 201 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
ProLLM - SQL Disambiguation53.1Score (%)100
OpenVLM MMMU - Management83.3Accuracy (%)99.6
OpenVLM MMBench V1.1 CN - Nature Relation96.7Accuracy (%)99.5
OpenVLM MMMU - Clinical Medicine80Accuracy (%)99.5
OpenVLM MMBench V1.1 CN - Spatial Relationship90.7Accuracy (%)99.3
OpenVLM MMMU - Art and Design77.5Accuracy (%)99.1
OpenVLM MMBench V1.1 EN - Spatial Relationship89.3Accuracy (%)98.9
OpenVLM MMMU - Business84Accuracy (%)98.9
OpenVLM AI2D - Rock Strata92.7Accuracy (%)98.6
OpenVLM MMBench V1.1 CN - Image Emotion87.8Accuracy (%)98.6
OpenVLM MMMU - Art86.7Accuracy (%)98.2
OpenVLM MMMU - Economics86.7Accuracy (%)98.1

Interactive version: theaggregate.ai/model?slug=gpt-4-5 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.