Claude Sonnet 5 vs GPT-5.6 Luna
Claude Sonnet 5 leads GPT-5.6 Luna by 1.0 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Sonnet 5 wins 11 and GPT-5.6 Luna wins 8 of 19 shared benchmarks. GPT-5.6 Luna is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Sonnet 5 | GPT-5.6 Luna |
|---|---|---|
| Agentic coding | 57 | 59 |
| Long-horizon autonomy | 65 | 61 |
| Knowledge-work agents | 43 | 39 |
| Computer use | 40 | 40 |
| Security & context learning | 51 | 51 |
Shared benchmarks
| Benchmark | Claude Sonnet 5 | GPT-5.6 Luna |
|---|---|---|
| APEX-Agentsindex | 54.5% | 43.0% |
| DeepSWEindex | 53.8% | 67.2% |
| Vending-Bench 2index | $6,378 | $4,095 |
| CursorBenchindex | 34.1% | 35.9% |
| Chess Puzzles | 45.0% | 43.0% |
| CritPt | 16.9% | 20.6% |
| DTBench | 92.5% | 89.1% |
| FrontierCode | 42.7% | 39.8% |
| FrontierMath Tier 4 | 29.3% | 61.0% |
| FrontierMath Tiers 1-3 | 65.6% | 82.1% |
| GPQA Diamond | 90.5% | 91.6% |
| LMCA | 50.0% | 48.5% |
| MirrorCode | 63.3% | 21.1% |
| OTIS Mock AIME | 94.7% | 98.3% |
| ProofBench | 77.0% | 60.0% |
| SciCode | 54.3% | 53.6% |
| SimpleBench | 60.6% | 46.8% |
| SimpleQA Verified | 33.7% | 41.0% |
| WebDev Arena | 1537 | 1519 |
Specs and price
| Claude Sonnet 5 | GPT-5.6 Luna | |
|---|---|---|
| Agent Index | 52.4 (#32) | 51.4 (#34) |
| Agent $/M | $0.92 | $0.10 |
| Input / output $/M | $2.00 / $10.00 | $0.20 / $1.20 |
| Context | 1M | 1.1M |
| Released | Jun 30, 2026 | Jul 9, 2026 |
| Weights | Closed | Closed |
| API hosts | 5 | 5 |
Don’t pick one. Route.
Shogo can run Claude Sonnet 5 and GPT-5.6 Luna side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo