KAT Coder Pro V2: benchmark results
Kwaipilot's flagship agentic-coding KAT-Coder successor (MoE, 72B active). Provider: Kuaishou. Released 2026-03-27. Access: API.
Aggregate Rating 1683 ± 46, rank #239 of 1656 rated models, from 43 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA Terminal-Bench Hard | 49.24 | Accuracy (%) | 94.5 |
| Tinybird AI SQL Benchmark - First-Attempt Success Rate | 100 | Questions answered with a valid query on the first attempt ( | 94 |
| AA TAU-2 Bench | 89.47 | Accuracy (%) | 83.9 |
| Tau2-Bench Telecom | 89.47 | Success Rate (%) | 83.8 |
| Tinybird AI SQL Benchmark - Exactness | 53.71 | Result exactness vs human reference queries (0-100) | 79.6 |
| AA IFBench | 66.67 | Accuracy (%) | 79.5 |
| AA GPQA Diamond | 85.45 | Accuracy (%) | 78.8 |
| SnakeBench | 27 | TrueSkill Rating | 76.8 |
| AA-Omniscience Hallucination Rate | 55.98 | Hallucination Rate (%) | 74.4 |
| Tinybird AI SQL Benchmark - Success Rate | 100 | Questions answered with a valid query within 3 retries (%) | 71.7 |
| Wolfram LLM Benchmarking Project | 53.8 | Correct Functionality (%) | 71.4 |
| AA-Omniscience Index - Software Engineering (SWE) - Dart | 0 | Omniscience Index | 71 |
Interactive version: theaggregate.ai/model?slug=kat-coder-pro-v2 · How It Works · Data refreshed daily, snapshot 2026-10-20.