Model benchmarks
Qwen3.6-35B-A3B-oQ4e-mtp local LLM performance
As of September 2026, Qwen3.6-35B-A3B-oQ4e-mtp runs at up to 85.8 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35B
Peak speed
85.8 tok/s
Average speed
81.4 tok/s
Avg PP
582.7 tok/s
Min memory
20.9 GB
Max context
262.144 tokens
Avg output / run
15.422 tokens
Avg runtime / run
3m 25s
Avg quality
75.1
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 | 70.8 | 75.1 | 79.4 |
| Agent Workflow | 75.4 | 77.4 | 79.3 |
| Code Generation | 53.3 | 64.3 | 75.2 |
| Role Play & Narrative | 78.7 | 83.4 | 88.0 |
| Research & Analysis | 71.8 | 75.5 | 79.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3.6-35B-A3B-oQ4e-mtp has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Apple M4 Pro | oMLX | — | 85.8 tok/s | 81.4 tok/s | 20.9 GB | 262.144 tokens | 75.1 | 2 |
Benchmark runs
All 2 Qwen3.6-35B-A3B-oQ4e-mtp runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Qwen3.6-35B-A3B-oQ4e-mtp good for coding?
- In our benchmarks, Qwen3.6-35B-A3B-oQ4e-mtp scores 64.3/100 for coding. It runs at about 81.4 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
- Is Qwen3.6-35B-A3B-oQ4e-mtp good for agentic (tool-using) tasks?
- In our benchmarks, Qwen3.6-35B-A3B-oQ4e-mtp scores 77.4/100 for agentic workflows. It runs at about 81.4 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
- How fast is Qwen3.6-35B-A3B-oQ4e-mtp for local inference?
- Across 2 community benchmark runs, Qwen3.6-35B-A3B-oQ4e-mtp reaches up to 85.8 tok/s and averages 81.4 tok/s, with the fastest results on Apple M4 Pro.
- How much memory does Qwen3.6-35B-A3B-oQ4e-mtp need?
- The leanest observed configuration used about 20.9 GB of memory.
- Which tools have been used to run Qwen3.6-35B-A3B-oQ4e-mtp?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.