Claude Opus 5 vs Kimi K3
Claude Opus 5 leads Kimi K3 by 10.5 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Opus 5 wins 22 and Kimi K3 wins 1 of 23 shared benchmarks. Kimi K3 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Opus 5 | Kimi K3 |
|---|---|---|
| Agentic coding | 74 | 63 |
| Long-horizon autonomy | 74 | 65 |
| Knowledge-work agents | 52 | 42 |
| Computer use | 56 | 43 |
| Security & context learning | 59 | 53 |
Shared benchmarks
| Benchmark | Claude Opus 5 | Kimi K3 |
|---|---|---|
| APEX-Agentsindex | 65.8% | 47.8% |
| DeepSWEindex | 73.6% | 68.5% |
| Vending-Bench 2index | $11,182 | $5,165 |
| FrontierSWEindex | 52.0% | 25.9% |
| PostTrainBenchindex | 35.7% | 31.5% |
| ARC-AGI | 97.5% | 94.5% |
| ARC-AGI-2 | 90.4% | 60.4% |
| Chess Puzzles | 43.0% | 37.0% |
| CritPt | 29.1% | 23.4% |
| DTBench | 97.9% | 91.2% |
| FrontierCode | 53.4% | 44.2% |
| FrontierMath Tier 4 | 73.2% | 39.0% |
| FrontierMath Tiers 1-3 | 85.6% | 72.2% |
| GPQA Diamond | 93.9% | 93.1% |
| LMCA | 64.5% | 52.7% |
| MirrorCode | 87.8% | 54.4% |
| OTIS Mock AIME | 98.9% | 97.2% |
| ProofBench | 99.0% | 87.0% |
| SciCode | 56.4% | 59.5% |
| SimpleBench | 80.6% | 60.7% |
| SimpleQA Verified | 59.9% | 50.6% |
| WebDev Arena | 1687 | 1674 |
| WeirdML v3 | 19.0% | 7.5% |
Specs and price
| Claude Opus 5 | Kimi K3 | |
|---|---|---|
| Agent Index | 65.4 (#11) | 54.9 (#26) |
| Agent $/M | $2.31 | $1.39 |
| Input / output $/M | $5.00 / $25.00 | $3.00 / $15.00 |
| Context | 1M | 1M |
| Released | Jul 24, 2026 | Jul 16, 2026 |
| Weights | Closed | Open |
| API hosts | 6 | 29 |
Don’t pick one. Route.
Shogo can run Claude Opus 5 and Kimi K3 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo