GPT-5: benchmark results
OpenAI's standard GPT-5 model for broad reasoning, coding, writing, and tool-use tasks. Provider: OpenAI. Released 2025-08-07. Access: API.
Aggregate Rating 1763 ± 2, rank #98 of 1656 rated models, from 4536 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| 360CityArena - Landmark Search (Image) | 48 | Success rate (%; the agent must stop within 10 m of a landma | 100 |
| 360CityArena - Relational Spatial Reasoning | 32 | Accuracy (%; the agent names the landmark that stands in a g | 100 |
| A2RBench - Semantic Tasks | 52 | Accuracy (%) on the 352 semantic (knowledge-dependent) rule | 100 |
| A2RBench - Symbolic Tasks (Original Symbols) | 41.3 | Accuracy (%) on the 351 symbolic (structure-based) rule task | 100 |
| AA MATH-500 | 99.4 | Accuracy (%) | 100 |
| ALBA | 91 | Mean judge score (0-100) over ALBA's 800 expert-written open | 100 |
| ALBA - Culture-bound Semantics | 88.3 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
| ALBA - Discourse Analysis | 98.5 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
| ALBA - Language Variety | 98.3 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
| ALBA - Morphology | 98.5 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
| ALBA - Phonetics and Phonology | 86.3 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
| ALBA - Word Plays | 91.5 | Mean judge score (0-100) on ALBA's 100 expert-written pt-PT | 100 |
Interactive version: theaggregate.ai/model?slug=gpt-5 · How It Works · Data refreshed daily, snapshot 2026-10-20.