Model comparison
Claude Opus 5 vs GPT-5.6 Luna
Claude Opus 5 is the stronger model overall, scoring 67.8 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 22× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.
Last verified . 50 shared benchmarks.
Summary
- They share 50 benchmarks with published results for both. Claude Opus 5 scores higher in 10 categories and GPT-5.6 Luna in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Opus 5 leads 77.2 to 47.6.
- The biggest single-benchmark swing is ProofBench: 99% for Claude Opus 5 and 60% for GPT-5.6 Luna.
- GPT-5.6 Luna is cheaper at $0.20 / $1.20 per million input/output tokens, against $5 / $25 for Claude Opus 5.
- GPT-5.6 Luna accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Opus 5 | GPT-5.6 Luna | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 67.8 | 54.6 |
| Released | 2026-07-24 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $5 | $0.20 |
| Output $ / M tokens | $25 | $1.20 |
| Results tracked | 57 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 5 leads
Claude Opus 5: 67.5 (#5), GPT-5.6 Luna: 54.5 (#28)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| DeepSWE | 73.6% | 67.2% |
| FrontierCode | 53.4% | 39.8% |
| CursorBench | 46.6% | 35.9% |
| LMArena WebDev | 1691 | 1519 |
| SciCode | 56.4% | 53.6% |
| WeirdML | 91.8% | 60.9% |
| LMArena Coding | 1534 | 1466 |
| ALE-Bench | 2,165 | 1,667 |
| FrontierSWE | 52% | — |
Agentic & Tool Use Claude Opus 5 leads
Claude Opus 5: 55.6 (#1), GPT-5.6 Luna: 34.4 (#45)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| APEX-Agents | 65.8% | 43% |
| BALROG | 63.4% | 45.6% |
| GDP.pdf | 24% | 22.7% |
| Vending-Bench 2 | 11,182 | 4,095 |
| OSWorld 2.0 | 31.4% | — |
| τ²-bench Banking | 48.7% | — |
| PostTrainBench | 35% | — |
| GBAEval | 79.6% | — |
Reasoning Claude Opus 5 leads
Claude Opus 5: 77.2 (#4), GPT-5.6 Luna: 47.6 (#43)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| ARC-AGI-2 | 90.4% | 59.5% |
| SimpleBench | 80.6% | 46.8% |
| NYT Connections (extended) | 94.3% | 69.4% |
| ARC-AGI-1 | 97.5% | 88% |
| CritPt | 29.1% | 20.6% |
| Chess Puzzles | 42% | 40% |
| LMArena Hard Prompts | 1526 | 1451 |
| Mystery Game Puzzles | 59% | 21% |
| DTBench | 97.9% | 89.1% |
| LMCA | 64.5% | 48.5% |
| Epoch Capabilities Index | 162.78 | 156.39 |
| Kagi LLM Benchmark | — | 49.1% |
| EBR-Bench | 45.7% | — |
| Surface Evolver Bench | — | 61.9% |
| Bench to the Future 3 | 0.12 | — |
Math Claude Opus 5 leads
Claude Opus 5: 86.2 (#8), GPT-5.6 Luna: 77.7 (#14)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| FrontierMath (Tiers 1-3) | 85.6% | 82.1% |
| FrontierMath Tier 4 | 73.2% | 61% |
| OTIS Mock AIME 2024-2025 | 98.9% | 98.3% |
| ProofBench | 99% | 60% |
| LMArena Math | 1531 | 1458 |
Knowledge Claude Opus 5 leads
Claude Opus 5: 66.8 (#9), GPT-5.6 Luna: 58.5 (#34)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| GPQA Diamond | 93.9% | 91.6% |
| SimpleQA Verified | 59.9% | 41% |
| LMArena Expert | 1557 | 1478 |
Multimodal Claude Opus 5 leads
Claude Opus 5: 50.8 (#8), GPT-5.6 Luna: 42.7 (#28)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Vision | 1319 | 1258 |
| Blueprint-Bench 2 | 30.4% | 22.6% |
| Furniture Assembly | 60.8% | 42.5% |
| LMArena Document | 1516 | 1457 |
Multilingual Claude Opus 5 leads
Claude Opus 5: 58.8 (#4), GPT-5.6 Luna: 52.8 (#78)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Non-English | 1501 | 1417 |
| LMArena Chinese | 1574 | 1470 |
| LMArena French | 1519 | 1456 |
| LMArena German | 1524 | 1454 |
| LMArena Japanese | 1516 | 1411 |
| LMArena Korean | 1521 | 1415 |
| LMArena Russian | 1507 | 1428 |
| LMArena Spanish | 1519 | 1448 |
Instruction Following Claude Opus 5 leads
Claude Opus 5: 79.2 (#7), GPT-5.6 Luna: 75.6 (#57)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Instruction Following | 1517 | 1437 |
Long Context Claude Opus 5 leads
Claude Opus 5: 46.5 (#21), GPT-5.6 Luna: 43.9 (#82)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Longer Query | 1515 | 1436 |
Writing & Preference Claude Opus 5 leads
Claude Opus 5: 79.2 (#1), GPT-5.6 Luna: 68.0 (#29)
| Benchmark | Claude Opus 5 | GPT-5.6 Luna |
|---|---|---|
| LMArena Text | 1507 | 1431 |
| LMArena Creative Writing | 1491 | 1396 |
| EQ-Bench Creative Writing | 2133 | 1829 |
| EQ-Bench 4 | 1385 | 1156 |
| LMArena Multi-Turn | 1499 | 1434 |
Frequently asked questions
Is Claude Opus 5 better than GPT-5.6 Luna?
Claude Opus 5 is the stronger model overall, scoring 67.8 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 22× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 5 or GPT-5.6 Luna?
GPT-5.6 Luna is cheaper. It lists at $0.20 per million input tokens and $1.20 per million output tokens; Claude Opus 5 lists at $5 and $25.
Is Claude Opus 5 or GPT-5.6 Luna better for coding?
Claude Opus 5 scores higher on coding benchmarks: 67.5 versus 54.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Luna does, with 1.05M tokens against 1M.
How many benchmarks do Claude Opus 5 and GPT-5.6 Luna share?
50 benchmarks have published results for both models. Claude Opus 5 has 57 scored results on Noometry and GPT-5.6 Luna has 52.