Leaderboard / Model
Claude Opus 4.7
Anthropic · Closed weights · Released Apr 16, 2026 · $5 in · $25 out per 1M tokens
59Overall · rank 12
Score by job
Coding
88 · rank 3 · 4 of 5 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| Terminal-Bench | 80.2% | claude-opus-4-7_unknown | 0.91 |
| FrontierCode | 38.5% | claude-opus-4-7_unknown | -0.11 |
| SWE-bench Verified | 83.5% | claude-opus-4-7_max | 2.14 |
| LMArena Coding | 1519 rating | claude-opus-4-7-high | 1.38 |
Agents
41 · rank 10 · 2 of 4 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| APEX-Agents | 49.2% | claude-opus-4-7_max | -0.30 |
| OSWorld 2.0 | 18.2% | claude-opus-4-7_max | 0.05 |
Reasoning
45 · rank 24 · 4 of 4 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| GPQA Diamond | 90.2% | claude-opus-4-7_xhigh | -0.57 |
| ARC-AGI-2 | 75.8% | claude-opus-4-7_max | 0.55 |
| SimpleBench | 61.7% | claude-opus-4-7_unknown | -0.37 |
| Humanity’s Last Exam | 36.2% | claude-opus-4-7_unknown | -0.19 |
Math
52 · rank 16 · 5 of 5 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| FrontierMath Tiers 1–3 | 70.2% | claude-opus-4-7_max | 0.15 |
| FrontierMath Tier 4 | 31.7% | claude-opus-4-7_max | -0.39 |
| OTIS Mock AIME | 97.8% | claude-opus-4-7_xhigh | 0.55 |
| ProofBench | 54.0% | claude-opus-4-7_max | 0.25 |
| LMArena Math | 1498 rating | claude-opus-4-7-high | 0.66 |
Writing
85 · rank 6 · 2 of 2 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| LMArena Creative Writing | 1485 rating | claude-opus-4-7-high | 1.02 |
| LMArena Instruction Following | 1498 rating | claude-opus-4-7-high | 1.24 |
Design
41 · rank 20 · 1 of 1 benchmarks
| Benchmark | Result | Variant | z |
|---|---|---|---|
| LMArena WebDev | 1557 rating | claude-opus-4-7 | -0.01 |