ChatGPT 5.1
OpenAI·
LLMsflagshipcloud
Overview
Major update with "Instant" and "Thinking" modes for adaptive reasoning.
Capabilities and innovations
Adaptive reasoningInstant/Thinking modesPersonalizationNode-based Agent Builder
Benchmarks
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 86.6% | unsourced |
| Humanity’s Last Exam (no tools) | 35.4% | unsourced |
| SWE-bench Verified | 67.8% | unsourced |
| MMLU | 92.3% | unsourced |
- SWE-bench Pro: not reported by the vendor
- MMLU-Pro: not reported by the vendor
- Terminal-Bench 2.x: benchmark did not exist at release
- MMMU-Pro: benchmark did not exist at release
Reliability
- Hallucination rate (Vectara HHEM)
- 12.1% (lower is better)independent
- Agentic tool use (τ-bench)
- 74.2% (higher is better)third party
Links
More from OpenAI
Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.