Model benchmarks
qwen3.5-9b-sushi-coder-rl local LLM performance
As of May 2026, qwen3.5-9b-sushi-coder-rl runs at up to 35.7 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
9B
Peak speed
35.7 tok/s
Average speed
34.4 tok/s
Min memory
n/a
Max context
8,192 tokens
Avg quality
54.8
Benchmark runs
2
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 2 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 | 48.0 | 54.8 | 61.6 |
| Agent Workflow | 67.2 | 71.6 | 76.0 |
| Code Generation | 47.1 | 50.7 | 54.3 |
| Role Play & Narrative | 7.0 | 39.0 | 70.9 |
| Research & Analysis | 54.1 | 58.0 | 61.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.5-9b-sushi-coder-rl has been benchmarked on, ranked by peak token generation speed. Last updated May 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3060 Ti | LM Studio | — | 35.7 tok/s | 34.4 tok/s | n/a | 8,192 tokens | 54.8 | 2 |
Benchmark runs
All 2 qwen3.5-9b-sushi-coder-rl runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.5-9b-sushi-coder-rl good for coding?
- In our benchmarks, qwen3.5-9b-sushi-coder-rl scores 50.7/100 for coding. It runs at about 34.4 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is qwen3.5-9b-sushi-coder-rl good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.5-9b-sushi-coder-rl scores 71.6/100 for agentic workflows. It runs at about 34.4 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is qwen3.5-9b-sushi-coder-rl for local inference?
- Across 2 community benchmark runs, qwen3.5-9b-sushi-coder-rl reaches up to 35.7 tok/s and averages 34.4 tok/s, with the fastest results on NVIDIA GeForce RTX 3060 Ti.
- Which tools have been used to run qwen3.5-9b-sushi-coder-rl?
- Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.