# Claude Opus 5 vs GPT-5 Mini

> Claude Opus 5 is the stronger model overall, scoring 67.7 to 41.7 on the Noometry Index. GPT-5 Mini costs 15× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-opus-5-vs-gpt-5-mini
- Last updated: 2026-10-10
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. Claude Opus 5 scores higher in 10 categories and GPT-5 Mini in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Opus 5 leads 77.1 to 24.1.
- The biggest single-benchmark swing is ProofBench: 99% for Claude Opus 5 and 9% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $5 / $25 for Claude Opus 5.
- Claude Opus 5 accepts more context: 1M tokens versus 400K.

## Snapshot

| | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 67.7 | 41.7 |
| Rank | 4 | 115 |
| Context | 1M | 400K |
| Input $/M | $5 | $0.25 |
| Output $/M | $25 | $2 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Opus 5: 67.0 (#5)
- GPT-5 Mini: 40.0 (#134)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| SciCode | 56.4% | 39.2% |
| WeirdML | 91.8% | 52.7% |
| LMArena Coding | 1533 | 1406 |
| ALE-Bench | 2,165 | 799.77 |
| SWE-bench Verified | — | 64.7% |
| DeepSWE | 73.6% | — |
| FrontierCode | 53.4% | — |
| SWE-bench Verified (bash only) | — | 59.8% |
| CursorBench | 46.6% | — |
| LMArena WebDev | 1694 | — |
| SWE-bench Multilingual | — | 39.7% |
| FrontierSWE | 52% | — |
| AlgoTune | — | 1.38 |

## Agentic & Tool Use

- Claude Opus 5: 55.5 (#1)
- GPT-5 Mini: 31.0 (#64)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| Vending-Bench 2 | 11,182 | -31.18 |
| Terminal-Bench | — | 34.8% |
| APEX-Agents | 65.8% | — |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| OSWorld 2.0 | 31.4% | — |
| τ²-bench Banking | 48.7% | — |
| PostTrainBench | 35% | — |
| BALROG | 63.4% | — |
| GBAEval | 79.6% | — |
| GDP.pdf | 24% | — |

## Reasoning

- Claude Opus 5: 77.1 (#3)
- GPT-5 Mini: 24.1 (#153)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 90.4% | 4.4% |
| ARC-AGI-1 | 97.5% | 54.3% |
| CritPt | 29.1% | 0% |
| Chess Puzzles | 42% | 30% |
| LMArena Hard Prompts | 1525 | 1380 |
| Mystery Game Puzzles | 59% | 10% |
| DTBench | 97.9% | 80.5% |
| LMCA | 64.5% | 34.2% |
| Epoch Capabilities Index | 162.67 | 145.51 |
| SimpleBench | 80.6% | — |
| Kagi LLM Benchmark | — | 70.3% |
| NYT Connections (extended) | 94.3% | — |
| EnigmaEval | — | 8.2% |
| EBR-Bench | 45.7% | — |
| Bench to the Future 3 | 0.12 | — |
| ForecastBench | — | 61 |

## Math

- Claude Opus 5: 86.1 (#6)
- GPT-5 Mini: 46.3 (#59)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 85.6% | 46.7% |
| FrontierMath Tier 4 | 73.2% | 12.2% |
| OTIS Mock AIME 2024-2025 | 98.9% | 86.7% |
| ProofBench | 99% | 9% |
| LMArena Math | 1533 | 1377 |
| Omni-MATH | — | 72.2% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 27.2% |
| FrontierMath Tier 4 (v1) | — | 6.3% |

## Knowledge

- Claude Opus 5: 66.8 (#8)
- GPT-5 Mini: 45.5 (#80)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 93.9% | 75% |
| SimpleQA Verified | 59.9% | 21.6% |
| LMArena Expert | 1558 | 1377 |
| Humanity's Last Exam | — | 19.4% |
| MMLU-Pro | — | 83.5% |
| Confabulations | — | 13.3% |
| Vectara Hallucination Rate | — | 12.9% |
| GPQA (HELM) | — | 75.6% |

## Multimodal

- Claude Opus 5: 51.1 (#6)
- GPT-5 Mini: 35.0 (#80)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1321 | 1202 |
| VPCT | — | 40.2% |
| Blueprint-Bench 2 | 30.4% | — |
| Furniture Assembly | 60.8% | — |
| LMArena Document | 1516 | — |

## Multilingual

- Claude Opus 5: 58.8 (#3)
- GPT-5 Mini: 49.1 (#131)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1500 | 1363 |
| LMArena Chinese | 1575 | 1387 |
| LMArena French | 1522 | 1385 |
| LMArena German | 1522 | 1366 |
| LMArena Japanese | 1510 | 1342 |
| LMArena Korean | 1526 | 1306 |
| LMArena Russian | 1507 | 1362 |
| LMArena Spanish | 1519 | 1355 |

## Instruction Following

- Claude Opus 5: 79.3 (#5)
- GPT-5 Mini: 76.4 (#43)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1516 | 1357 |
| IFEval | — | 92.7% |

## Long Context

- Claude Opus 5: 46.5 (#20)
- GPT-5 Mini: 41.9 (#127)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1516 | 1355 |
| Fiction.LiveBench | — | 69.4% |

## Writing & Preference

- Claude Opus 5: 79.3 (#2)
- GPT-5 Mini: 55.4 (#143)

| Benchmark | Claude Opus 5 | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1506 | 1373 |
| LMArena Creative Writing | 1491 | 1325 |
| EQ-Bench Creative Writing | 2133 | 1313 |
| LMArena Multi-Turn | 1496 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| WildBench | — | 85.5% |
| EQ-Bench 4 | 1385 | — |

## FAQ

### Is Claude Opus 5 better than GPT-5 Mini?

Claude Opus 5 is the stronger model overall, scoring 67.7 to 41.7 on the Noometry Index. GPT-5 Mini costs 15× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.

### Which is cheaper, Claude Opus 5 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 5 lists at $5 and $25.

### Is Claude Opus 5 or GPT-5 Mini better for coding?

Claude Opus 5 scores higher on coding benchmarks: 67.0 versus 40.0 in the Noometry coding category.

### Which has the bigger context window?

Claude Opus 5 does, with 1M tokens against 400K.

### How many benchmarks do Claude Opus 5 and GPT-5 Mini share?

37 benchmarks have published results for both models. Claude Opus 5 has 57 scored results on Noometry and GPT-5 Mini has 60.