Up to four models
GPT-5.5 vs GPT-5.4
On the index (GPT-4 = 100), GPT-5.5 scores 172 and GPT-5.4 166. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
GPT-5.5
172156–188
WeightsClosed
ReleasedApr 2026
Price in / out$5.0 / $30
Context1.05M
Core tests taken6 of 9
GPT-5.4
166150–182
WeightsClosed
ReleasedMar 2026
Price in / out$2.5 / $15
Context1.05M
Core tests taken7 of 9
GPT-5.5 leads GPT-5.4 by 6 points, ahead on 5 of 5 shared core tests, at 2.0× the price.
Test by testthird-party results
GPT-5.5GPT-5.4
Every result as a table
| Test | GPT-5.5 | GPT-5.4 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 94.0% | 93.3% |
| LiveBench Data Analysis | 81.6% | 79.3% |
| LiveBench IF | 70.7% | 70.2% |
| LiveBench Language | 87.4% | 82.6% |
| LiveBench Reasoning | 89.6% | 88.1% |
| GPQA Diamond (Vals AI) | 93.2% | 91.7% |
| MMLU Pro | 88.1% | 87.5% |
| MMMU Pro | 88.3% | 87.5% |
| EBR-bench | 34.3% | 25.4% |
| SimpleQA Verified | 63.0% | 45.1% |
| Maths | ||
| FrontierMath (tiers 1-3) | 85.3% | 78.6% |
| AIME-style maths (OTIS mock) | 100.0% | 97.8% |
| LiveBench Mathematics | 95.9% | 94.2% |
| Chess puzzles | 63.0% | 48.0% |
| FrontierMath (2025 set) | 51.7% | 47.6% |
| FrontierMath tier 4 (2025 set) | 35.4% | 27.1% |
| FrontierMath (tier 4) | 72.5% | 49.0% |
| Coding | ||
| SWE-bench Verified | 80.6% | 76.9% |
| LiveBench Agentic Coding | 54.0% | 53.8% |
| LiveBench Coding | 82.2% | 77.5% |
| Code Migration | 45.2% | 35.0% |
| LiveCodeBench | 85.3% | 84.1% |
| ProgramBench | 0.5% | 0.5% |
| SWE-bench | 82.6% | 78.2% |
| Vibe Code Bench v1.1 | 69.8% | 67.4% |
| DeepSWE | 67.0% | 51.8% |
| GSO-Bench | 40.2% | 31.4% |
| MirrorCode | 23.3% | 23.3% |
| WeirdML | 84.9% | 77.7% |
| Agents | ||
| APEX-Agents | 55.1% | 52.4% |
| SkillsBench | 62.2% | 51.7% |
| DTBench | 96.0% | 94.4% |
| DeepResearch Bench | 54.0% | 35.1% |
| Furniture assembly | 44.2% | 37.5% |
| LMCA | 54.3% | 52.0% |
| Mystery game puzzles | 56.0% | 37.0% |
| Professional work | ||
| CorpFin v2 | 68.4% | 65.3% |
| LegalBench | 86.5% | 86.0% |
| MedCode | 49.1% | 41.3% |
| MedScribe | 86.9% | 77.5% |
| MortgageTax | 68.8% | 68.3% |
| TaxEval v2 | 75.0% | 74.0% |
| Novel problems | ||
| ARC-AGI-2 | 85.0% | 74.0% |
| ARC-AGI-1 | 95.0% | 93.7% |
| ARC-AGI-3 | 0.4% | 0.2% |
Task length · METR45 sec
Tests at least two of these models have taken. Bold rows are the core tests.