Model benchmarks
Qwen3.6-35B-A3B-Q4_K_XL local LLM performance
As of August 2026, Qwen3.6-35B-A3B-Q4_K_XL runs at up to 159.0 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
159.0 tok/s
Average speed
156.9 tok/s
Min memory
17.1 GB
Max context
65,536 tokens
Avg output / run
17,702 tokens
Avg runtime / run
1m 55s
Avg quality
84.0
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 | 83.9 | 84.0 | 84.1 |
| Agent Workflow | 76.9 | 84.4 | 91.9 |
| Code Generation | 80.2 | 82.2 | 84.2 |
| Role Play & Narrative | 84.8 | 86.6 | 88.4 |
| Research & Analysis | 79.6 | 82.8 | 86.0 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3.6-35B-A3B-Q4_K_XL has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XT/7900 XTX/7900 GRE/7900M | llama.cpp | Q4_K | 159.0 tok/s | 156.9 tok/s | 17.1 GB | 65,536 tokens | 84.0 | 2 |
Frequently asked questions
- How fast is Qwen3.6-35B-A3B-Q4_K_XL for local inference?
- Across 2 community benchmark runs, Qwen3.6-35B-A3B-Q4_K_XL reaches up to 159.0 tok/s and averages 156.9 tok/s, with the fastest results on AMD Radeon RX 7900 XT/7900 XTX/7900 GRE/7900M.
- How much memory does Qwen3.6-35B-A3B-Q4_K_XL need?
- The leanest observed configuration used about 17.1 GB of memory (quantizations tested: Q4_K).
- Which tools have been used to run Qwen3.6-35B-A3B-Q4_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.