KAT-Coder-Pro V2.5 is a flagship agentic coding model designed to take on complete issues or end-to-end business workflows, autonomously locate and modify code in real repositories, run tests, and deliver production-ready results. Trained with large-scale agentic reinforcement learning and a multi-expert architecture, it combines strong long-horizon engineering execution with the frontend aesthetic generation capabilities introduced in V2. It is well suited for real-repository software development, workflow automation such as data analysis and report generation, and seamless integration with coding agents like Claude Code and OpenHands.
Back to Models
-$0.74/ M tokens$2.96/ M tokensRead:0.15/ M tokensWrite:-/ M tokens256K0.97s41.7tps
Providers
Route requests across multiple providers. Copy a provider slug to set your preference.
ProviderCache Hit RateContext
StreamlakeSunset
streamlake
Uptime
24hoursDirect request success rate on AI Gateway and per-provider.
Aug 31, 05:04
Streamlake100.00%
Throughput
24hoursP50 throughput on live AI Gateway traffic, in tokens per second (TPS).
Aug 31, 01:00
Streamlake41.7 tps
Latency
24hoursP50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds.
Aug 31, 01:00
Streamlake0.97s
Activity
Token volume and request traffic to this model over time.
Token Consumption
Aug 30
Streamlake888
Token volume broken down by the provider that served each request. Larger segments indicate providers handling more of this model's traffic.
Benchmarks
Scores on standardized evaluations. Higher percentages are better — and rank percentile shows
Metrics sourced fromArtificial Analysis
Artificial AnalysisNo benchmark data is available for this model.
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for. View All
Token Ranking
No app usage is available for this period.