Model benchmarks
Kwaipilot-KAT-Coder_t1p95k20m0 local LLM performance
As of November 2026, Kwaipilot-KAT-Coder_t1p95k20m0 runs at up to 101.2 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
101.2 tok/s
Average speed
100.5 tok/s
Avg PP
496.9 tok/s
Min memory
47.2 GB
Max context
32,768 tokens
Avg output / run
10,908 tokens
Avg runtime / run
1m 54s
Avg quality
79.2
Benchmark runs
3
GPUs tested
1
Quality by task
Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 3 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 76.6 | 79.2 | 82.3 |
| Agent Workflow | 63.5 | 75.5 | 88.8 |
| Code Generation | 71.9 | 74.6 | 77.2 |
| Role Play & Narrative | 81.6 | 83.4 | 84.5 |
| Research & Analysis | 80.9 | 83.1 | 84.5 |
Performance by hardware and tool
Every hardware/tool/quantization combination Kwaipilot-KAT-Coder_t1p95k20m0 has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Tesla V100-PCIE-32GB | llama.cpp | — | 101.2 tok/s | 100.5 tok/s | 47.2 GB | 32,768 tokens | 79.2 | 3 |
Benchmark runs
All 3 Kwaipilot-KAT-Coder_t1p95k20m0 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Kwaipilot-KAT-Coder_t1p95k20m0 good for coding?
- In our benchmarks, Kwaipilot-KAT-Coder_t1p95k20m0 scores 74.6/100 for coding. It runs at about 100.5 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is Kwaipilot-KAT-Coder_t1p95k20m0 good for agentic (tool-using) tasks?
- In our benchmarks, Kwaipilot-KAT-Coder_t1p95k20m0 scores 75.5/100 for agentic workflows. It runs at about 100.5 tok/s, so if you want more speed, swift-1.5-qwen3.8-flash-next-iq3_s is faster (~259.0 tok/s) and still scores well for agentic workflows (88.9/100).
- How fast is Kwaipilot-KAT-Coder_t1p95k20m0 for local inference?
- Across 3 community benchmark runs, Kwaipilot-KAT-Coder_t1p95k20m0 reaches up to 101.2 tok/s and averages 100.5 tok/s, with the fastest results on Tesla V100-PCIE-32GB.
- How much memory does Kwaipilot-KAT-Coder_t1p95k20m0 need?
- The leanest observed configuration used about 47.2 GB of memory.
- Which tools have been used to run Kwaipilot-KAT-Coder_t1p95k20m0?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.