Models/KAT-Coder-Pro V1

MDL

KAT-Coder-Pro V1

KwaiKAT

Verified against Artificial Analysis · Jul 13, 2026

Capability

percentile index

Price /M

$0.3 / $1.2

in / out

Context

—

Avg score

63.7

10 benchmarks

AA Intelligence

34.6

index

Speed

109 tok/s

TTFT 0.9s

Released

Nov 11, 2025

About

KAT-Coder-Pro V1 is a proprietary model in the KAT family from KwaiKAT. Benchmarked on 10 evals, averaging 63.7.

Benchmark scores · 10

avg 63.7 — higher is better

AIMEartificial-analysis · 2026-07

94.7

τ²-benchartificial-analysis · 2026-07

88.6

MMLU-Proartificial-analysis · 2026-07

81.3

GPQA Diamondartificial-analysis · 2026-07

76.4

LiveCodeBenchartificial-analysis · 2026-07

74.7

AA Long-Context Reasoningartificial-analysis · 2026-07

IFBenchartificial-analysis · 2026-07

68.4

SciCodeartificial-analysis · 2026-07

36.6

Humanity's Last Examartificial-analysis · 2026-07

33.4

Terminal-Bench Hardartificial-analysis · 2026-07

9.1

Details

PublisherKwaiKAT

FamilyKAT

Parameters—

License—

WeightsClosed

ReleasedNov 11, 2025

Performance & price

Output speed109 tok/s

Latency (TTFT)0.9s

Price in / out$0.3 / $1.2 per 1M

Blended price$0.525 / 1M

Coding index58.9

Math index94.7

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$1.2/M