KwaiKAT
KAT Coder Pro V2
KAT Coder Pro V2 is a KwaiKAT language-model profile in the Easy Benchmarks snapshot. Use this page to compare its measured Artificial Analysis scores, output speed, time to first token, pricing, and relative ranking against other models in the local catalog.
KwaiKAT model releasesOperational Metrics
Output Speed110.7 tok/s
First Token1.89s
Blended Price$0.525/M
Model Metadata
Queryable facts extracted from the upstream model payload.
ReleaseMar 27, 2026
Context Windown/a
Modalitiesn/a
Strength: Coding
Rank #24 across 410 models.
45.6
Strength: Overall
Rank #48 across 500 models.
43.8
Strength: GPQA
Rank #48 across 478 models.
85.5%
Watch Area: TTFT
Rank #224 across 293 models.
1.89s
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 43.8 | #48 |
| Artificial Analysis Coding Index | coding | 45.6 | #24 |
| GPQA | reasoning | 85.5% | #48 |
| Humanity's Last Exam | reasoning | 16.0% | #89 |
| SciCode |