Model comparison
Claude Opus 4 vs GPT-5 Mini
Claude Opus 4 is the stronger model overall, scoring 43.1 to 41.8 on the Noometry Index. GPT-5 Mini costs 44× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Last verified . 48 shared benchmarks.
Summary
- They share 48 benchmarks with published results for both. Claude Opus 4 scores higher in 5 categories and GPT-5 Mini in 5 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Claude Opus 4 leads 47.2 to 40.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 64.4% for Claude Opus 4 and 86.7% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $15 / $75 for Claude Opus 4.
- GPT-5 Mini accepts more context: 400K tokens versus 200K.
Side by side
| Claude Opus 4 | GPT-5 Mini | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 43.1 | 41.8 |
| Released | 2025-05-22 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 400K |
| Max output | 32K | 128K |
| Input $ / M tokens | $15 | $0.25 |
| Output $ / M tokens | $75 | $2 |
| Results tracked | 56 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4 leads
Claude Opus 4: 47.2 (#62), GPT-5 Mini: 40.1 (#146)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| SWE-bench Verified | 70.7% | 64.7% |
| SWE-bench Verified (bash only) | 67.6% | 59.8% |
| WeirdML | 43.7% | 52.7% |
| LMArena Coding | 1442 | 1406 |
| AlgoTune | 1.33 | 1.38 |
| Aider Polyglot | 72% | — |
| SWE-bench Multilingual | — | 39.7% |
| SciCode | — | 39.2% |
| GSO | 6.9% | — |
| ALE-Bench | — | 799.77 |
Agentic & Tool Use Claude Opus 4 leads
Claude Opus 4: 34.8 (#42), GPT-5 Mini: 31.1 (#70)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | — | 34.8% |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| Cybench | 38% | — |
| DeepResearch Bench | 46.8% | — |
| LMArena Search | 1127 | — |
| METR Time Horizons | 63.9% | — |
| Vending-Bench 2 | — | -31.18 |
Reasoning Claude Opus 4 leads
Claude Opus 4: 27.3 (#121), GPT-5 Mini: 23.9 (#168)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 8.6% | 4.4% |
| Kagi LLM Benchmark | 74.3% | 70.3% |
| ARC-AGI-1 | 35.7% | 54.3% |
| CritPt | 0.3% | 0% |
| EnigmaEval | 5.6% | 8.2% |
| LMArena Hard Prompts | 1399 | 1380 |
| DTBench | 81.6% | 80.5% |
| LMCA | 37.4% | 34.2% |
| Epoch Capabilities Index | 142.67 | 145.52 |
| ForecastBench | 61.1 | 61 |
| SimpleBench | 58.8% | — |
| Chess Puzzles | — | 30% |
| Mystery Game Puzzles | — | 10% |
Math GPT-5 Mini leads
Claude Opus 4: 42.0 (#86), GPT-5 Mini: 46.7 (#69)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 86.7% |
| Omni-MATH | 61.6% | 72.2% |
| LMArena Math | 1390 | 1378 |
| MATH Level 5 | 85% | 97.8% |
| FrontierMath (Feb 2025 set) | 4.5% | 27.2% |
| FrontierMath Tier 4 (v1) | 4.2% | 6.3% |
| FrontierMath (Tiers 1-3) | — | 46.7% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 9% |
Knowledge GPT-5 Mini leads
Claude Opus 4: 44.0 (#88), GPT-5 Mini: 45.6 (#86)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 76.3% | 75% |
| Humanity's Last Exam | 10.7% | 19.4% |
| MMLU-Pro | 87.5% | 83.5% |
| Confabulations | 15.9% | 13.3% |
| Vectara Hallucination Rate | 12% | 12.9% |
| GPQA (HELM) | 70.8% | 75.6% |
| LMArena Expert | 1386 | 1379 |
| SimpleQA Verified | — | 21.6% |
Multimodal GPT-5 Mini leads
Claude Opus 4: 31.5 (#106), GPT-5 Mini: 35.6 (#85)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1192 | 1202 |
| VPCT | 38% | 40.2% |
| GeoBench | 49% | — |
Multilingual Too close to call
Claude Opus 4: 48.8 (#138), GPT-5 Mini: 48.9 (#137)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1362 | 1363 |
| LMArena Chinese | 1386 | 1385 |
| LMArena French | 1372 | 1386 |
| LMArena German | 1391 | 1366 |
| LMArena Japanese | 1331 | 1341 |
| LMArena Korean | 1321 | 1308 |
| LMArena Russian | 1392 | 1362 |
| LMArena Spanish | 1389 | 1355 |
Instruction Following Too close to call
Claude Opus 4: 77.1 (#28), GPT-5 Mini: 76.2 (#46)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| IFEval | 91.8% | 92.7% |
| LMArena Instruction Following | 1406 | 1357 |
Long Context GPT-5 Mini leads
Claude Opus 4: 39.6 (#172), GPT-5 Mini: 41.9 (#132)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| Fiction.LiveBench | 61.1% | 69.4% |
| LMArena Longer Query | 1422 | 1355 |
Writing & Preference Claude Opus 4 leads
Claude Opus 4: 61.2 (#89), GPT-5 Mini: 55.2 (#148)
| Benchmark | Claude Opus 4 | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1377 | 1373 |
| LMArena Creative Writing | 1387 | 1325 |
| Short-Story Creative Writing | 83.6% | 83.1% |
| EQ-Bench Creative Writing | 1580 | 1313 |
| WildBench | 85.2% | 85.5% |
| LMArena Multi-Turn | 1396 | 1363 |
Frequently asked questions
Is Claude Opus 4 better than GPT-5 Mini?
Claude Opus 4 is the stronger model overall, scoring 43.1 to 41.8 on the Noometry Index. GPT-5 Mini costs 44× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4 or GPT-5 Mini?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or GPT-5 Mini better for coding?
Claude Opus 4 scores higher on coding benchmarks: 47.2 versus 40.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 200K.
How many benchmarks do Claude Opus 4 and GPT-5 Mini share?
48 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and GPT-5 Mini has 60.