KAT-Coder-Pro V2.5: benchmark results
Provider: Kuaishou. Released 2026-07-10. Access: API.
Aggregate Rating 1700 ± 22, rank #348 of 10916 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Wolfram LLM Benchmarking Project | 55.5 | Correct Functionality (%) | 73.2 |
| AI Chess Leaderboard (Reasoning) | 718 | Elo | 52.1 |
| ObviousBench | 90.97 | Answer pass³ (%) | 46.6 |
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 92 | Questions answered with a valid query on the first attempt ( | 42.2 |
| Tinybird AI SQL Benchmark - Success Rate | 98 | Questions answered with a valid query within 3 retries (%) | 35.1 |
| SvelteBench | 82.2 | Average pass@1 (%) | 33.4 |
| KiloBench - Terminal Bench 2.0 | 50.34 | Task completion in Kilo's agent harness, mean of 5 attempts | 33.3 |
| Tinybird AI SQL Benchmark - Exactness | 43.99 | Result exactness vs human reference queries (0-100) | 28.8 |
| AI Chess Leaderboard (Continuation) | 466 | Elo | 24.9 |
| Gert Labs Rankings | 31.82 | GScore (%) | 21.4 |
Interactive version: theaggregate.ai/model?slug=kat-coder-pro-v2-5 · How It Works · Data refreshed daily, snapshot 2026-11-08.