Model comparison
Claude Opus 4.8 vs Claude Opus 5
Claude Opus 5 is the stronger model overall, scoring 67.8 to 60.7 on the Noometry Index.
Last verified . 54 shared benchmarks.
Summary
- They share 54 benchmarks with published results for both. Claude Opus 4.8 scores higher in 0 categories and Claude Opus 5 in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Opus 5 leads 77.2 to 64.7.
- The biggest single-benchmark swing is ProofBench: 69% for Claude Opus 4.8 and 99% for Claude Opus 5.
- Both cost about the same: $5 input and $25 output per million tokens.
Side by side
| Claude Opus 4.8 | Claude Opus 5 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 60.7 | 67.8 |
| Released | 2026-05-28 | 2026-07-24 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1M |
| Max output | 128K | 128K |
| Input $ / M tokens | $5 | $5 |
| Output $ / M tokens | $25 | $25 |
| Results tracked | 65 | 57 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 5 leads
Claude Opus 4.8: 59.9 (#12), Claude Opus 5: 67.5 (#5)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| DeepSWE | 59% | 73.6% |
| FrontierCode | 46.5% | 53.4% |
| LMArena WebDev | 1556 | 1691 |
| SciCode | 53.5% | 56.4% |
| WeirdML | 82.9% | 91.8% |
| LMArena Coding | 1490 | 1534 |
| ALE-Bench | 1,564 | 2,165 |
| CursorBench | — | 46.6% |
| FrontierSWE | — | 52% |
| GSO | 47.1% | — |
Agentic & Tool Use Claude Opus 5 leads
Claude Opus 4.8: 47.6 (#11), Claude Opus 5: 55.6 (#1)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| APEX-Agents | 48.9% | 65.8% |
| OSWorld 2.0 | 20.6% | 31.4% |
| τ²-bench Banking | 39.7% | 48.7% |
| PostTrainBench | 33.8% | 35% |
| GBAEval | 70.9% | 79.6% |
| GDP.pdf | 24% | 24% |
| Vending-Bench 2 | 5,787 | 11,182 |
| Remote Labor Index | 8.3% | — |
| DeepResearch Bench | 50.2% | — |
| BALROG | — | 63.4% |
| LMArena Search | 1204 | — |
Reasoning Claude Opus 5 leads
Claude Opus 4.8: 64.7 (#16), Claude Opus 5: 77.2 (#4)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| ARC-AGI-2 | 72.1% | 90.4% |
| SimpleBench | 64.8% | 80.6% |
| NYT Connections (extended) | 91.1% | 94.3% |
| ARC-AGI-1 | 92.5% | 97.5% |
| CritPt | 20.9% | 29.1% |
| Chess Puzzles | 34% | 42% |
| EBR-Bench | 28.6% | 45.7% |
| LMArena Hard Prompts | 1482 | 1526 |
| Mystery Game Puzzles | 36% | 59% |
| DTBench | 94.9% | 97.9% |
| LMCA | 57.5% | 64.5% |
| Bench to the Future 3 | 0.14 | 0.12 |
| Epoch Capabilities Index | 158.21 | 162.78 |
| Kagi LLM Benchmark | 88.8% | — |
| EnigmaEval | 23.5% | — |
| Surface Evolver Bench | 87.5% | — |
| ForecastBench | 59.9 | — |
Math Claude Opus 5 leads
Claude Opus 4.8: 78.4 (#13), Claude Opus 5: 86.2 (#8)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 80% | 85.6% |
| FrontierMath Tier 4 | 56.1% | 73.2% |
| OTIS Mock AIME 2024-2025 | 98.3% | 98.9% |
| ProofBench | 69% | 99% |
| LMArena Math | 1487 | 1531 |
| MathArena Final-Answer Competitions | 91.8% | — |
| FrontierMath (Feb 2025 set) | 47.2% | — |
| FrontierMath Tier 4 (v1) | 31.3% | — |
Knowledge Claude Opus 5 leads
Claude Opus 4.8: 61.3 (#29), Claude Opus 5: 66.8 (#9)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| GPQA Diamond | 91% | 93.9% |
| SimpleQA Verified | 53% | 59.9% |
| LMArena Expert | 1502 | 1557 |
Multimodal Claude Opus 5 leads
Claude Opus 4.8: 42.9 (#26), Claude Opus 5: 50.8 (#8)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| LMArena Vision | 1294 | 1319 |
| Blueprint-Bench 2 | 14.5% | 30.4% |
| Furniture Assembly | 42.5% | 60.8% |
| LMArena Document | 1475 | 1516 |
Multilingual Claude Opus 5 leads
Claude Opus 4.8: 55.2 (#33), Claude Opus 5: 58.8 (#4)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| LMArena Non-English | 1450 | 1501 |
| LMArena Chinese | 1507 | 1574 |
| LMArena French | 1481 | 1519 |
| LMArena German | 1472 | 1524 |
| LMArena Japanese | 1440 | 1516 |
| LMArena Korean | 1432 | 1521 |
| LMArena Russian | 1474 | 1507 |
| LMArena Spanish | 1466 | 1519 |
Instruction Following Claude Opus 5 leads
Claude Opus 4.8: 77.4 (#24), Claude Opus 5: 79.2 (#7)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| LMArena Instruction Following | 1476 | 1517 |
Long Context Claude Opus 5 leads
Claude Opus 4.8: 45.4 (#35), Claude Opus 5: 46.5 (#21)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| LMArena Longer Query | 1483 | 1515 |
Writing & Preference Claude Opus 5 leads
Claude Opus 4.8: 72.0 (#16), Claude Opus 5: 79.2 (#1)
| Benchmark | Claude Opus 4.8 | Claude Opus 5 |
|---|---|---|
| LMArena Text | 1461 | 1507 |
| LMArena Creative Writing | 1454 | 1491 |
| EQ-Bench Creative Writing | 1840 | 2133 |
| EQ-Bench 4 | 1281 | 1385 |
| LMArena Multi-Turn | 1476 | 1499 |
Frequently asked questions
Is Claude Opus 4.8 better than Claude Opus 5?
Claude Opus 5 is the stronger model overall, scoring 67.8 to 60.7 on the Noometry Index.
Which is cheaper, Claude Opus 4.8 or Claude Opus 5?
Claude Opus 5 is cheaper. It lists at $5 per million input tokens and $25 per million output tokens; Claude Opus 4.8 lists at $5 and $25.
Is Claude Opus 4.8 or Claude Opus 5 better for coding?
Claude Opus 5 scores higher on coding benchmarks: 67.5 versus 59.9 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Claude Opus 4.8 and Claude Opus 5 share?
54 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and Claude Opus 5 has 57.