KwaiKATKAT-Coder-Pro V2
KAT-Coder-Pro V2 is a KwaiKAT proprietary model, released 27 Mar 2026. Its best configuration ranks #319 of 785 on the BenchLeader Index at 52.4 ±10.4, in the upper half. The ± is the point: 535 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (66) and lowest in reasoning (41). Last measured 27 Mar 2026.stale: no new result in six months
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- 256k
- Full answer
- –
- Cost per run
- –
- Released
- 27 Mar 2026
- Overall index52
- Reasoning41
- Agents & tools47
- Knowledge51
- Instruction following66
- Long context61
- Composite54
How its index has moved
9 Sept 2026 to 21 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 85.5%#125 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 16.1%#244 | Artificial Analysis | |
| CritPt | 0.0%#315 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Terminal-Bench 2.1 (AA)not in index | 70.0%#72 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Terminal-Bench Hard | 49.2%#24 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 89.5%#67 | Artificial Analysis | |
| GDPval-AA v2.1not in index | 11.7%#111 | Artificial Analysis | |
| τ³-Banking (AA) | 4.5%#191 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -22.5#252 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 21.5%#283 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 44.0%#146 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFBench | 66.7%#92 | Artificial Analysis |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 73.0%#182 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 21.7#210 | Artificial Analysis |
See also
Questions people ask
- What are KAT-Coder-Pro V2's benchmark scores?
- As of 25 Oct 2026, KAT-Coder-Pro V2 scores 52.4 ±10.4 on the BenchLeader Index, strongest in instruction following (66) and weakest in reasoning (41). That rests on 7 independent benchmark results, and every score on this page links to its publisher.
- How does KAT-Coder-Pro V2 rank?
- KAT-Coder-Pro V2 is #319 of 785 ranked configurations on 25 Oct 2026. 535 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- When was KAT-Coder-Pro V2 released?
- KwaiKAT released KAT-Coder-Pro V2 on 27 Mar 2026.
Data as of 25 Oct 2026. Compare with another model.
Cite as: BenchLeader, “KAT-Coder-Pro V2: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/kat-coder-pro-v2.