Model comparison
Claude Opus 4.8 vs Kimi K2.6
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 47.7 on the Noometry Index. Kimi K2.6 costs 5.8× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.
Last verified . 48 shared benchmarks.
Summary
- They share 48 benchmarks with published results for both. Claude Opus 4.8 scores higher in 10 categories and Kimi K2.6 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Claude Opus 4.8 leads 47.6 to 21.9.
- The biggest single-benchmark swing is GBAEval: 70.9% for Claude Opus 4.8 and 0.9% for Kimi K2.6.
- Kimi K2.6 is cheaper at $0.95 / $4 per million input/output tokens, against $5 / $25 for Claude Opus 4.8.
- Claude Opus 4.8 accepts more context: 1M tokens versus 262K.
- Kimi K2.6 has downloadable open weights; the other is API-only.
Side by side
| Claude Opus 4.8 | Kimi K2.6 | |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 60.7 | 47.7 |
| Released | 2026-05-28 | 2026-04-20 |
| Weights | Proprietary | Open |
| Context window | 1M | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $5 | $0.95 |
| Output $ / M tokens | $25 | $4 |
| Results tracked | 65 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4.8 leads
Claude Opus 4.8: 59.9 (#12), Kimi K2.6: 50.7 (#43)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena WebDev | 1556 | 1509 |
| SciCode | 53.5% | 53.5% |
| WeirdML | 82.9% | 55.9% |
| LMArena Coding | 1490 | 1488 |
| ALE-Bench | 1,564 | 1,093 |
| SWE-bench Verified | — | 76.7% |
| DeepSWE | 59% | — |
| FrontierCode | 46.5% | — |
| GSO | 47.1% | — |
Agentic & Tool Use Claude Opus 4.8 leads
Claude Opus 4.8: 47.6 (#11), Kimi K2.6: 21.9 (#137)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| OSWorld 2.0 | 20.6% | 4.6% |
| GBAEval | 70.9% | 0.9% |
| GDP.pdf | 24% | 12% |
| Vending-Bench 2 | 5,787 | 6,205 |
| APEX-Agents | 48.9% | — |
| Remote Labor Index | 8.3% | — |
| τ²-bench Banking | 39.7% | — |
| DeepResearch Bench | 50.2% | — |
| PostTrainBench | 33.8% | — |
| ExploitBench | — | 18.4% |
| LMArena Search | 1204 | — |
Reasoning Claude Opus 4.8 leads
Claude Opus 4.8: 64.7 (#16), Kimi K2.6: 40.5 (#55)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| NYT Connections (extended) | 91.1% | 87.2% |
| CritPt | 20.9% | 8% |
| Chess Puzzles | 34% | 26% |
| EBR-Bench | 28.6% | 2.4% |
| LMArena Hard Prompts | 1482 | 1470 |
| Mystery Game Puzzles | 36% | 18% |
| DTBench | 94.9% | 90.9% |
| LMCA | 57.5% | 37.3% |
| Epoch Capabilities Index | 158.21 | 151.05 |
| ARC-AGI-2 | 72.1% | — |
| SimpleBench | 64.8% | — |
| Kagi LLM Benchmark | 88.8% | — |
| ARC-AGI-1 | 92.5% | — |
| EnigmaEval | 23.5% | — |
| Surface Evolver Bench | 87.5% | — |
| Bench to the Future 3 | 0.14 | — |
| ForecastBench | 59.9 | — |
Math Claude Opus 4.8 leads
Claude Opus 4.8: 78.4 (#13), Kimi K2.6: 57.0 (#41)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 80% | 57.2% |
| FrontierMath Tier 4 | 56.1% | 25.6% |
| MathArena Final-Answer Competitions | 91.8% | 72.9% |
| OTIS Mock AIME 2024-2025 | 98.3% | 96.1% |
| ProofBench | 69% | 16% |
| LMArena Math | 1487 | 1475 |
| FrontierMath (Feb 2025 set) | 47.2% | 39% |
| FrontierMath Tier 4 (v1) | 31.3% | 14.6% |
Knowledge Claude Opus 4.8 leads
Claude Opus 4.8: 61.3 (#29), Kimi K2.6: 54.0 (#54)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| GPQA Diamond | 91% | 90.8% |
| SimpleQA Verified | 53% | 34.9% |
| LMArena Expert | 1502 | 1491 |
| Vectara Hallucination Rate | — | 10.8% |
Multimodal Claude Opus 4.8 leads
Claude Opus 4.8: 42.9 (#26), Kimi K2.6: 31.6 (#103)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena Vision | 1294 | 1283 |
| Blueprint-Bench 2 | 14.5% | 3.9% |
| Furniture Assembly | 42.5% | 21.7% |
| LMArena Document | 1475 | 1451 |
Multilingual Too close to call
Claude Opus 4.8: 55.2 (#33), Kimi K2.6: 54.9 (#37)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena Non-English | 1450 | 1446 |
| LMArena Chinese | 1507 | 1521 |
| LMArena French | 1481 | 1471 |
| LMArena German | 1472 | 1450 |
| LMArena Japanese | 1440 | 1443 |
| LMArena Korean | 1432 | 1427 |
| LMArena Russian | 1474 | 1446 |
| LMArena Spanish | 1466 | 1464 |
Instruction Following Claude Opus 4.8 leads
Claude Opus 4.8: 77.4 (#24), Kimi K2.6: 76.3 (#43)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena Instruction Following | 1476 | 1451 |
Long Context Too close to call
Claude Opus 4.8: 45.4 (#35), Kimi K2.6: 44.9 (#52)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena Longer Query | 1483 | 1468 |
Writing & Preference Claude Opus 4.8 leads
Claude Opus 4.8: 72.0 (#16), Kimi K2.6: 68.5 (#26)
| Benchmark | Claude Opus 4.8 | Kimi K2.6 |
|---|---|---|
| LMArena Text | 1461 | 1455 |
| LMArena Creative Writing | 1454 | 1434 |
| EQ-Bench Creative Writing | 1840 | 1725 |
| EQ-Bench 4 | 1281 | 1202 |
| LMArena Multi-Turn | 1476 | 1453 |
Frequently asked questions
Is Claude Opus 4.8 better than Kimi K2.6?
Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 47.7 on the Noometry Index. Kimi K2.6 costs 5.8× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4.8 or Kimi K2.6?
Kimi K2.6 is cheaper. It lists at $0.95 per million input tokens and $4 per million output tokens; Claude Opus 4.8 lists at $5 and $25.
Is Claude Opus 4.8 or Kimi K2.6 better for coding?
Claude Opus 4.8 scores higher on coding benchmarks: 59.9 versus 50.7 in the Noometry coding category.
Which has the bigger context window?
Claude Opus 4.8 does, with 1M tokens against 262K.
How many benchmarks do Claude Opus 4.8 and Kimi K2.6 share?
48 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and Kimi K2.6 has 51.