Model comparison
Claude Opus 4 vs GPT-5
GPT-5 is the stronger model overall, scoring 50.9 to 43.1 on the Noometry Index.
Last verified . 55 shared benchmarks.
Summary
- They share 55 benchmarks with published results for both. Claude Opus 4 scores higher in 2 categories and GPT-5 in 8 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 39.6.
- The biggest single-benchmark swing is Fiction.LiveBench: 61.1% for Claude Opus 4 and 97.2% for GPT-5.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $15 / $75 for Claude Opus 4.
- GPT-5 accepts more context: 400K tokens versus 200K.
Side by side
| Claude Opus 4 | GPT-5 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 43.1 | 50.9 |
| Released | 2025-05-22 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 400K |
| Max output | 32K | 128K |
| Input $ / M tokens | $15 | $1.25 |
| Output $ / M tokens | $75 | $10 |
| Results tracked | 56 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
Claude Opus 4: 47.2 (#62), GPT-5: 50.3 (#47)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| SWE-bench Verified | 70.7% | 73.6% |
| SWE-bench Verified (bash only) | 67.6% | 65% |
| Aider Polyglot | 72% | 88% |
| GSO | 6.9% | 6.9% |
| WeirdML | 43.7% | 60.7% |
| LMArena Coding | 1442 | 1436 |
| AlgoTune | 1.33 | 1.67 |
| LMArena WebDev | — | 1418 |
| SciCode | — | 42.9% |
| ALE-Bench | — | 1,162 |
Agentic & Tool Use Claude Opus 4 leads
Claude Opus 4: 34.8 (#42), GPT-5: 33.1 (#56)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| DeepResearch Bench | 46.8% | 49.6% |
| LMArena Search | 1127 | 1133 |
| METR Time Horizons | 63.9% | 69.6% |
| Terminal-Bench | — | 49.6% |
| GDPval | — | 34.8% |
| Remote Labor Index | — | 1.7% |
| Cybench | 38% | — |
| BALROG | — | 32.8% |
Reasoning GPT-5 leads
Claude Opus 4: 27.3 (#121), GPT-5: 38.3 (#64)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 8.6% | 9.9% |
| SimpleBench | 58.8% | 56.7% |
| Kagi LLM Benchmark | 74.3% | 72.7% |
| ARC-AGI-1 | 35.7% | 65.7% |
| CritPt | 0.3% | 12.6% |
| EnigmaEval | 5.6% | 10.5% |
| LMArena Hard Prompts | 1399 | 1416 |
| DTBench | 81.6% | 90.7% |
| LMCA | 37.4% | 40% |
| Epoch Capabilities Index | 142.67 | 150 |
| ForecastBench | 61.1 | 61.4 |
| Chess Puzzles | — | 37% |
| EBR-Bench | — | 12.7% |
| Mystery Game Puzzles | — | 23% |
Math GPT-5 leads
Claude Opus 4: 42.0 (#86), GPT-5: 55.0 (#44)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 91.4% |
| Omni-MATH | 61.6% | 64.7% |
| LMArena Math | 1390 | 1407 |
| MATH Level 5 | 85% | 98.1% |
| FrontierMath (Feb 2025 set) | 4.5% | 32.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 12.5% |
| FrontierMath (Tiers 1-3) | — | 55.4% |
| FrontierMath Tier 4 | — | 22% |
| ProofBench | — | 18% |
Knowledge GPT-5 leads
Claude Opus 4: 44.0 (#88), GPT-5: 56.6 (#43)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| GPQA Diamond | 76.3% | 86.2% |
| Humanity's Last Exam | 10.7% | 25.3% |
| MMLU-Pro | 87.5% | 86.3% |
| Confabulations | 15.9% | 10.3% |
| Vectara Hallucination Rate | 12% | 14.7% |
| GPQA (HELM) | 70.8% | 79.2% |
| LMArena Expert | 1386 | 1419 |
| SimpleQA Verified | — | 50.1% |
Multimodal GPT-5 leads
Claude Opus 4: 31.5 (#106), GPT-5: 46.8 (#13)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| LMArena Vision | 1192 | 1232 |
| GeoBench | 49% | 81% |
| VPCT | 38% | 66% |
Multilingual GPT-5 leads
Claude Opus 4: 48.8 (#138), GPT-5: 51.4 (#110)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| LMArena Non-English | 1362 | 1397 |
| LMArena Chinese | 1386 | 1422 |
| LMArena French | 1372 | 1410 |
| LMArena German | 1391 | 1416 |
| LMArena Japanese | 1331 | 1409 |
| LMArena Korean | 1321 | 1360 |
| LMArena Russian | 1392 | 1406 |
| LMArena Spanish | 1389 | 1399 |
Instruction Following Claude Opus 4 leads
Claude Opus 4: 77.1 (#28), GPT-5: 73.8 (#113)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| IFEval | 91.8% | 87.5% |
| LMArena Instruction Following | 1406 | 1388 |
Long Context GPT-5 leads
Claude Opus 4: 39.6 (#172), GPT-5: 69.5 (#2)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| Fiction.LiveBench | 61.1% | 97.2% |
| LMArena Longer Query | 1422 | 1399 |
Writing & Preference GPT-5 leads
Claude Opus 4: 61.2 (#89), GPT-5: 63.4 (#65)
| Benchmark | Claude Opus 4 | GPT-5 |
|---|---|---|
| LMArena Text | 1377 | 1406 |
| LMArena Creative Writing | 1387 | 1365 |
| Short-Story Creative Writing | 83.6% | 86% |
| EQ-Bench Creative Writing | 1580 | 1627 |
| WildBench | 85.2% | 85.7% |
| LMArena Multi-Turn | 1396 | 1426 |
Frequently asked questions
Is Claude Opus 4 better than GPT-5?
GPT-5 is the stronger model overall, scoring 50.9 to 43.1 on the Noometry Index.
Which is cheaper, Claude Opus 4 or GPT-5?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or GPT-5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 47.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5 does, with 400K tokens against 200K.
How many benchmarks do Claude Opus 4 and GPT-5 share?
55 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and GPT-5 has 69.