GPT-5.1vsGPT-5
Across 13 shared benchmarks, GPT-5.1 leads overall: GPT-5.1 wins 7, GPT-5 wins 5, with 1 ties and an average score difference of +1.12.
GPT-5.17 wins(54%)Ties1(38%)5 winsGPT-5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 13 shared benchmarks.
General Knowledge
GPT-5.1 3/4| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| HLE | 26.5083 / 157 | 35.2060 / 157 | -8.70 |
| ARC-AGI-2 | 17.6033 / 59 | 9.9037 / 59 | +7.70 |
| ARC-AGI | 72.8025 / 65 | 65.7030 / 65 | +7.10 |
| GPQA Diamond | 88.1028 / 178 | 87.3037 / 178 | +0.80 |
Math and Reasoning
GPT-5 2/4| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| AIME2025 | 9428 / 106 | 99.609 / 106 | -5.60 |
| Simple Bench | 53.2010 / 27 | 56.708 / 27 | -3.50 |
| FrontierMath | 26.7013 / 60Thinking High (With Tools) | 24.8015 / 60 | +1.90 |
| FrontierMath - Tier 4 | 12.5029 / 80Thinking High (With Tools) | 12.5029 / 80Thinking High (No Tools) | — |
Coding and Software Engineer
GPT-5.1 2/2| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 50.8030 / 43Thinking High (No Tools) | 36.3041 / 43 | +14.50 |
| SWE-bench Verified | 76.3030 / 108 | 72.8046 / 108 | +3.50 |
Agent Level Benchmark
GPT-5 1/1| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| τ²-Bench - Telecom | 95.6014 / 35Thinking High (With Tools) | 95.8013 / 35 | -0.20 |
AI Agent - Information Search
GPT-5 1/1| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| BrowseComp | 50.8036 / 45Thinking High (No Tools) | 54.9032 / 45 | -4.10 |
Multimodal Understanding
GPT-5.1 1/1| Benchmark | GPT-5.1 | GPT-5 | Diff |
|---|---|---|---|
| MMMU | 85.402 / 28 | 84.205 / 28 | +1.20 |
Specs
| Field | GPT-5.1 | GPT-5 |
|---|---|---|
| Publisher | OpenAI | OpenAI |
| Release date | 2025-11-12 | 2025-08-07 |
| Model type | Reasoning model | Foundation model |
| Architecture | Dense | Dense |
| Parameters | Not available | Not available |
| Context length | 400K | 400K |
| Max output | 128K | 128K |
Summary
- GPT-5.1leads in:General Knowledge (3/4), Coding and Software Engineer (2/2), Multimodal Understanding (1/1)
- GPT-5leads in:Math and Reasoning (2/4), Agent Level Benchmark (1/1), AI Agent - Information Search (1/1)
On average across the 13 shared benchmarks, GPT-5.1 scores 1.12 higher.
Largest single-benchmark gap: SWE-Bench Pro - Public — GPT-5.1 50.80 vs GPT-5 36.30 (+14.50).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.