About
KAT-Coder-Pro V2 is the latest high-performance model in KwaiKAT’s KAT-Coder series, designed for complex enterprise-grade software engineering and SaaS integration. It builds on the agentic coding strengths of earlier versions,...
Related Models
Pricing
Input
$0.30
per 1M tokens
Output
$1.20
per 1M tokens
Blended
$0.52
per 1M tokens
Cheaper than 54% of models. Median price is $0.68/1M tokens.
Cost Calculator
Tokens per day1M
100K100M
Daily
$0.52
Monthly
$15.75
vs. Similar Models
GLM-5.1 (Non-reasoning)Q:0.0
$2.15+310%
GPT-5.5 (Non-reasoning)Q:0.0
$11.25+2043%
Grok 4.3 (low)Q:0.0
$1.56+198%
MiMo-V2-OmniQ:-0.4
$0.80+52%
Performance
103
tokens/sec
Faster than 56% of models
0.88
seconds
Faster than 63% of models
0.88
seconds
Faster than 73% of models
Market Median
93 tok/s
11% faster
Median TTFT
1.13s
22% faster
Throughput/Dollar
197
tok/s per $/1M
Speed Comparison
GPT-5 mini (medium)
103 tok/s+0%
Qwen3 Omni 30B A3B (Reasoning)
104 tok/s+0%
DeepSeek: DeepSeek V4 Flash
104 tok/s+0%
Context Window
256K
tokens
Larger than 56% of models
Max Output
80K
tokens
31% of context
Benchmarks
MMLU-ProNot evaluated
GPQA Diamond
85.5%
HLE
16.0%
LiveCodeBenchNot evaluated
SciCode
38.3%
TerminalBench Hard
49.2%
MATH-500Not evaluated
AIMENot evaluated
AIME 2025Not evaluated
IFBench
66.7%
Long Context Recall
66.0%
Tau2
89.5%
Market AverageTop Score