B.O.B. Leaderboard
Anthropicreleased 24 Jul 2026
Claude Opus 5
#10 of 9535 of 46 benchmarksConsistent
84.9
B.O.B.
Category profile
- Real use (hard to game)2×84
- Reasoning & knowledge76
- Math82
- Coding84
- Agents & real work91
- Independent indexes93
Reality check
Lab tests
84
Real use (9 tests)
84
Lab tests and real-use tests agree. Its lab-test and real-use scores are within 2 points of each other.
What counts as real usePrice per 1M tokens (OpenRouter)
$5.00 in · $25.00 out
Cost per task on the AA index
—
Output speed (Artificial Analysis)
—
§Receipts
Every benchmark result
“Beats” = the share of tracked models this one beat on that test. The raw score is its best reported setting.
| Benchmark | Raw | Beats | Source |
|---|---|---|---|
| Real use (hard to game) · Real use 84 | |||
| SimpleBenchas claude-opus-5_unknown | 80.6% | 97%of 30 | Epoch AI Benchmarking Hub |
| EQ-Bench Creative Writingas claude-opus-5 | 2133 | 96%of 46 | Sam Paech (open source) |
| SWE-rebenchas Opus 5 [high] | 63.4% | 94%of 34 | Nebius |
| LMArena WebDevas claude-opus-5-max | 1690 | 94%of 82 | LMArena (blind human votes) |
| Vals Indexas anthropic/claude-opus-5 | 63.7% | 88%of 42 | Vals AI (private test sets) |
| LMArena Textas claude-opus-5-high | 1490 | 86%of 80 | LMArena (blind human votes) |
| LMArena Agent Arenaas Claude Opus 5 (Max) | 0.092 | 86%of 50 | LMArena (real agent sessions) |
| LiveBench 2.0 (real-work suites)as claude-opus-5 | 81.8% | 84%of 39 | LiveBench / Abacus.AI |
| BridgeBench (overall)as Claude Opus 5 | 512 | 29%of 29 | BridgeMind |
| Reasoning & knowledge · Reasoning 76 | |||
| Mystery Game Puzzlesas claude-opus-5_max | 59.0% | 91%of 57 | Epoch AI Benchmarking Hub |
| ARC-AGI-2as claude-opus-5_max | 90.4% | 91%of 44 | Epoch AI Benchmarking Hub |
| CritPt (physics research)as claude-opus-5_max | 29.1% | 89%of 89 | Epoch AI Benchmarking Hub |
| GPQA Diamondas claude-opus-5_max | 93.9% | 84%of 70 | Epoch AI Benchmarking Hub |
| SimpleQA Verifiedas claude-opus-5_max | 59.9% | 73%of 57 | Epoch AI Benchmarking Hub |
| Chess Puzzlesas claude-opus-5_max | 43.0% | 63%of 73 | Epoch AI Benchmarking Hub |
| Blueprint-Bench 2as claude-opus-5_unknown | 30.4% | 43%of 31 | Epoch AI Benchmarking Hub |
| Math · Math 82 | |||
| ProofBenchas claude-opus-5_max | 99.0% | 93%of 68 | Epoch AI Benchmarking Hub |
| FrontierMath Tiers 1-3as claude-opus-5_max | 85.6% | 81%of 60 | Epoch AI Benchmarking Hub |
| OTIS Mock AIMEas claude-opus-5_max | 98.9% | 79%of 73 | Epoch AI Benchmarking Hub |
| FrontierMath Tier 4as claude-opus-5_max | 73.2% | 77%of 52 | Epoch AI Benchmarking Hub |
| Coding · Coding 84 | |||
| FrontierCodeas claude-opus-5_max | 53.4% | 96%of 45 | Epoch AI Benchmarking Hub |
| ALE-Benchas claude-opus-5_high | 2164.6% | 94%of 71 | Epoch AI Benchmarking Hub |
| WeirdMLas claude-opus-5_max | 91.8% | 92%of 52 | Epoch AI Benchmarking Hub |
| DeepSWEas claude-opus-5_max | 73.6% | 92%of 26 | Epoch AI Benchmarking Hub |
| SciCodeas claude-opus-5_max | 56.4% | 74%of 86 | Epoch AI Benchmarking Hub |
| FrontierSWEas claude-opus-5_max | 52.0% | 74%of 20 | Epoch AI Benchmarking Hub |
| CursorBenchas claude-opus-5_max | 46.6% | 69%of 17 | Epoch AI Benchmarking Hub |
| Agents & real work · Agents 91 | |||
| OSWorld 2.0as claude-opus-5_max | 31.4% | 100%of 9 | Epoch AI Benchmarking Hub |
| Game Boy Advance evalas claude-opus-5_unknown | 79.6% | 100%of 23 | Epoch AI Benchmarking Hub |
| DTBenchas claude-opus-5_xhigh | 97.9% | 97%of 67 | Epoch AI Benchmarking Hub |
| LMCAas claude-opus-5_xhigh | 64.5% | 97%of 67 | Epoch AI Benchmarking Hub |
| Vending-Bench 2as claude-opus-5_unknown | 11181.9% | 94%of 47 | Epoch AI Benchmarking Hub |
| APEX-Agentsas claude-opus-5_max | 65.8% | 89%of 48 | Epoch AI Benchmarking Hub |
| GDP.pdfas claude-opus-5_max | 24.0% | 63%of 37 | Epoch AI Benchmarking Hub |
| Independent indexes · Indexes 93 | |||
| Epoch Capabilities Indexas Claude Opus 5 | 162.6 | 93%of 81 | Epoch AI |