Claude Opus 4.6 vs Claude Opus 4.7
Claude Opus 4.7 leads Claude Opus 4.6 by 1.4 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Opus 4.6 wins 4 and Claude Opus 4.7 wins 17 of 21 shared benchmarks. Claude Opus 4.6 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Opus 4.6 | Claude Opus 4.7 |
|---|---|---|
| Agentic coding | 57 | 60 |
| Long-horizon autonomy | 67 | 69 |
| Knowledge-work agents | 41 | 41 |
| Computer use | 39 | 44 |
| Security & context learning | 57 | 52 |
Shared benchmarks
| Benchmark | Claude Opus 4.6 | Claude Opus 4.7 |
|---|---|---|
| APEX-Agentsindex | 46.3% | 49.2% |
| SWE-bench Verifiedindex | 78.7% | 83.5% |
| Terminal-Bench 2.0index | 79.8% | 80.2% |
| Vending-Bench 2index | $8,018 | $10,937 |
| GSO-Benchindex | 41.2% | 44.1% |
| ARC-AGI | 94.0% | 93.5% |
| ARC-AGI-2 | 69.2% | 75.8% |
| Chess Puzzles | 21.0% | 41.0% |
| DTBench | 91.2% | 94.7% |
| FrontierCode | 26.6% | 38.5% |
| FrontierMath Tier 4 | 26.8% | 31.7% |
| FrontierMath Tiers 1-3 | 66.0% | 70.2% |
| GPQA Diamond | 90.5% | 90.2% |
| Humanity's Last Exam | 34.4% | 36.2% |
| LMCA | 55.8% | 52.2% |
| MirrorCode | 33.3% | 43.3% |
| OTIS Mock AIME | 94.4% | 97.8% |
| ProofBench | 50.0% | 54.0% |
| SimpleBench | 67.6% | 61.7% |
| SimpleQA Verified | 47.0% | 51.7% |
| WebDev Arena | 1556 | 1567 |
Specs and price
| Claude Opus 4.6 | Claude Opus 4.7 | |
|---|---|---|
| Agent Index | 53.2 (#31) | 54.6 (#29) |
| Agent $/M | $2.31 | $2.31 |
| Input / output $/M | $5.00 / $25.00 | $5.00 / $25.00 |
| Context | 1M | 1M |
| Released | Feb 5, 2026 | Apr 16, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 5 |
Don’t pick one. Route.
Shogo can run Claude Opus 4.6 and Claude Opus 4.7 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo