Model comparison
GPT-5 vs GPT-5.5
GPT-5.5 is the stronger model overall, scoring 63.4 to 50.9 on the Noometry Index. GPT-5 costs 3.3× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Last verified . 50 shared benchmarks.
Summary
- They share 50 benchmarks with published results for both. GPT-5 scores higher in 1 category and GPT-5.5 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.5 leads 72.8 to 38.3.
- The biggest single-benchmark swing is ARC-AGI-2: 9.9% for GPT-5 and 85% for GPT-5.5.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $5 / $30 for GPT-5.5.
- GPT-5.5 accepts more context: 1.05M tokens versus 400K.
Side by side
| GPT-5 | GPT-5.5 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 50.9 | 63.4 |
| Released | 2025-08-07 | 2026-04-23 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $1.25 | $5 |
| Output $ / M tokens | $10 | $30 |
| Results tracked | 69 | 71 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.5 leads
GPT-5: 50.3 (#47), GPT-5.5: 58.2 (#17)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| SWE-bench Verified | 73.6% | 80.6% |
| LMArena WebDev | 1418 | 1513 |
| SciCode | 42.9% | 56.1% |
| GSO | 6.9% | 40.2% |
| WeirdML | 60.7% | 84.9% |
| LMArena Coding | 1436 | 1494 |
| ALE-Bench | 1,162 | 1,943 |
| DeepSWE | — | 67% |
| FrontierCode | — | 43% |
| SWE-bench Verified (bash only) | 65% | — |
| Aider Polyglot | 88% | — |
| MirrorCode | — | 10% |
| AlgoTune | 1.67 | — |
Agentic & Tool Use GPT-5.5 leads
GPT-5: 33.1 (#56), GPT-5.5: 50.7 (#6)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| Terminal-Bench | 49.6% | 84.7% |
| Remote Labor Index | 1.7% | 6.3% |
| DeepResearch Bench | 49.6% | 54% |
| LMArena Search | 1133 | 1242 |
| APEX-Agents | — | 55.1% |
| OSWorld 2.0 | — | 13% |
| GDPval | 34.8% | — |
| τ²-bench Banking | — | 44.6% |
| PostTrainBench | — | 27.2% |
| BALROG | 32.8% | — |
| ExploitBench | — | 47.4% |
| GBAEval | — | 53.2% |
| GDP.pdf | — | 26% |
| METR Time Horizons | 69.6% | — |
| Vending-Bench 2 | — | 7,524 |
Reasoning GPT-5.5 leads
GPT-5: 38.3 (#64), GPT-5.5: 72.8 (#11)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| ARC-AGI-2 | 9.9% | 85% |
| SimpleBench | 56.7% | 69% |
| Kagi LLM Benchmark | 72.7% | 88.8% |
| ARC-AGI-1 | 65.7% | 95% |
| CritPt | 12.6% | 27.1% |
| Chess Puzzles | 37% | 54% |
| EBR-Bench | 12.7% | 34.3% |
| LMArena Hard Prompts | 1416 | 1489 |
| Mystery Game Puzzles | 23% | 56% |
| DTBench | 90.7% | 96% |
| LMCA | 40% | 54.3% |
| Epoch Capabilities Index | 150 | 159.1 |
| ForecastBench | 61.4 | 60.6 |
| NYT Connections (extended) | — | 96.2% |
| EnigmaEval | 10.5% | — |
| Surface Evolver Bench | — | 88.1% |
| Bench to the Future 3 | — | 0.14 |
Math GPT-5.5 leads
GPT-5: 55.0 (#44), GPT-5.5: 81.7 (#11)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 55.4% | 85.3% |
| FrontierMath Tier 4 | 22% | 72.5% |
| OTIS Mock AIME 2024-2025 | 91.4% | 100% |
| ProofBench | 18% | 50% |
| LMArena Math | 1407 | 1486 |
| FrontierMath (Feb 2025 set) | 32.4% | 51.7% |
| FrontierMath Tier 4 (v1) | 12.5% | 35.4% |
| MathArena Final-Answer Competitions | — | 94.3% |
| Omni-MATH | 64.7% | — |
| MATH Level 5 | 98.1% | — |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.5 leads
GPT-5: 56.6 (#43), GPT-5.5: 64.4 (#17)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| GPQA Diamond | 86.2% | 94% |
| SimpleQA Verified | 50.1% | 63% |
| Vectara Hallucination Rate | 14.7% | 9.3% |
| LMArena Expert | 1419 | 1508 |
| Humanity's Last Exam | 25.3% | — |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.3% | — |
| GPQA (HELM) | 79.2% | — |
Multimodal Too close to call
GPT-5: 46.8 (#13), GPT-5.5: 46.9 (#12)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| LMArena Vision | 1232 | 1297 |
| GeoBench | 81% | — |
| VPCT | 66% | — |
| Blueprint-Bench 2 | — | 36.2% |
| Furniture Assembly | — | 44.2% |
| LMArena Document | — | 1486 |
Multilingual GPT-5.5 leads
GPT-5: 51.4 (#110), GPT-5.5: 56.4 (#20)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| LMArena Non-English | 1397 | 1467 |
| LMArena Chinese | 1422 | 1533 |
| LMArena French | 1410 | 1486 |
| LMArena German | 1416 | 1480 |
| LMArena Japanese | 1409 | 1498 |
| LMArena Korean | 1360 | 1460 |
| LMArena Russian | 1406 | 1473 |
| LMArena Spanish | 1399 | 1468 |
Instruction Following GPT-5.5 leads
GPT-5: 73.8 (#113), GPT-5.5: 77.5 (#18)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| LMArena Instruction Following | 1388 | 1479 |
| IFEval | 87.5% | — |
Long Context GPT-5 leads
GPT-5: 69.5 (#2), GPT-5.5: 48.3 (#12)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| LMArena Longer Query | 1399 | 1484 |
| Fiction.LiveBench | 97.2% | — |
| CL-bench Life | — | 22.2% |
Writing & Preference GPT-5.5 leads
GPT-5: 63.4 (#65), GPT-5.5: 72.7 (#13)
| Benchmark | GPT-5 | GPT-5.5 |
|---|---|---|
| LMArena Text | 1406 | 1472 |
| LMArena Creative Writing | 1365 | 1455 |
| EQ-Bench Creative Writing | 1627 | 1844 |
| LMArena Multi-Turn | 1426 | 1476 |
| Short-Story Creative Writing | 86% | — |
| WildBench | 85.7% | — |
| EQ-Bench 4 | — | 1315 |
Frequently asked questions
Is GPT-5 better than GPT-5.5?
GPT-5.5 is the stronger model overall, scoring 63.4 to 50.9 on the Noometry Index. GPT-5 costs 3.3× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Which is cheaper, GPT-5 or GPT-5.5?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-5.5 lists at $5 and $30.
Is GPT-5 or GPT-5.5 better for coding?
GPT-5.5 scores higher on coding benchmarks: 58.2 versus 50.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5.5 does, with 1.05M tokens against 400K.
How many benchmarks do GPT-5 and GPT-5.5 share?
50 benchmarks have published results for both models. GPT-5 has 69 scored results on Noometry and GPT-5.5 has 71.