Anthropic: Claude Opus 5
anthropic/claude-opus-5
Context: 1,000,000In: $5.00/1MOut: $25.00/1MUpdated 2026-10-04
Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis...
Intelligence
50.8
Coding
78.0
Agentic
56.5
Metric profile
Trend over time
Classic evals (curated from model cards)
Anthropic Claude (Opus) — manually maintained, may lag releases
MMLU-Pro
77.5%
GPQA Diamond
78.6%
HumanEval
95%
MATH-500
96.8%
AIME 2025
32%
Design Arena (ELO by category)
| Category | ELO | Win rate | Rank |
|---|---|---|---|
| asciiart | 1363 | 68.2% | #1 |
| gamedev | 1350 | 59% | #5 |
| 3d | 1349 | 61.1% | #7 |
| uicomponent | 1349 | 60.4% | #5 |
| mobileapps | 1348 | 67.1% | #1 |
| dataviz | 1345 | 60.8% | #6 |
| codecategories | 1329 | 57.9% | #5 |
| website | 1315 | 56.1% | #7 |
| svg | 1314 | 58.8% | #3 |
| fullstack | 1305 | 57.3% | #4 |
| godotgamedev | 1279 | 54.2% | #3 |
| androidnative | 1269 | 55.9% | #3 |
| python-pptxslides | 1267 | 54.4% | #5 |
| agenticgamedev | 1261 | 53.8% | #2 |
| webapps | 1259 | 54.5% | #7 |