GPT-4o Mini: benchmark results
OpenAI's small, low-cost multimodal GPT-4o tier with a 128K context that replaced GPT-3.5 Turbo in ChatGPT (July 2024). Provider: OpenAI. Released 2024-07-18. Access: API.
Unified ELO 1502 ± 1, rank #681 of 1392 rated models, from 986 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Galileo Tool Tasks - xLAM Multiple Tool Multiple Call | 99 | Accuracy (%) | 100 |
| HalluWorld | 28.1 | Overall (self-reported) | 100 |
| MEGA-Bench Task - 2D Image Jigsaw Puzzle Easy | 38.1 | Task Score (%) | 100 |
| MEGA-Bench Task - Action Sequence | 85.7 | Task Score (%) | 100 |
| MEGA-Bench Task - Bridge Strategies Advanced | 65.7 | Task Score (%) | 100 |
| MEGA-Bench Task - Bridge Strategies World-Class | 51.4 | Task Score (%) | 100 |
| MEGA-Bench Task - Constrained Generation Contain Length | 86.7 | Task Score (%) | 100 |
| MEGA-Bench Task - Ishihara Test | 58.6 | Task Score (%) | 100 |
| MEGA-Bench Task - MMSoc Misinformation GossipCop | 92.9 | Task Score (%) | 100 |
| MEGA-Bench Task - OCR Math Equation | 78.6 | Task Score (%) | 100 |
| MEGA-Bench Task - Painting QA | 92.9 | Task Score (%) | 100 |
| MEGA-Bench Task - Poetry Limerick | 100 | Task Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=gpt-4o-mini · How It Works · Data refreshed daily, snapshot 2026-09-05.