Model benchmarks
qwen3.6:35b-a3b-mtp-q4_K_M local LLM performance
As of November 2026, qwen3.6:35b-a3b-mtp-q4_K_M runs at up to 45.3 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35.5B
Peak speed
45.3 tok/s
Average speed
44.1 tok/s
Avg PP
251.8 tok/s
Min memory
16.4 GB
Max context
32,768 tokens
Avg output / run
18,169 tokens
Avg runtime / run
7m 19s
Avg quality
71.9
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 | 65.2 | 71.9 | 78.6 |
| Agent Workflow | 37.5 | 57.6 | 77.8 |
| Code Generation | 65.9 | 70.8 | 75.7 |
| Role Play & Narrative | 83.4 | 87.6 | 91.9 |
| Research & Analysis | 64.1 | 71.4 | 78.6 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.6:35b-a3b-mtp-q4_K_M 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 |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7600 XT | Ollama | Q4_K_M | 45.3 tok/s | 44.1 tok/s | 16.4 GB | 32,768 tokens | 71.9 | 2 |
Benchmark runs
All 2 qwen3.6:35b-a3b-mtp-q4_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.6:35b-a3b-mtp-q4_K_M good for coding?
- In our benchmarks, qwen3.6:35b-a3b-mtp-q4_K_M scores 70.8/100 for coding. It runs at about 44.1 tok/s, so if you want more speed, nvidia/Qwen3.8-27B-NVFP4 is faster (~63.3 tok/s) and still scores well for coding (82.7/100).
- Is qwen3.6:35b-a3b-mtp-q4_K_M good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.6:35b-a3b-mtp-q4_K_M scores 57.6/100 for agentic workflows. It runs at about 44.1 tok/s, so if you want more speed, swift-iq2_xs is faster (~91.5 tok/s) and still scores well for agentic workflows (91.4/100).
- How fast is qwen3.6:35b-a3b-mtp-q4_K_M for local inference?
- Across 2 community benchmark runs, qwen3.6:35b-a3b-mtp-q4_K_M reaches up to 45.3 tok/s and averages 44.1 tok/s, with the fastest results on AMD Radeon RX 7600 XT.
- How much memory does qwen3.6:35b-a3b-mtp-q4_K_M need?
- The leanest observed configuration used about 16.4 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run qwen3.6:35b-a3b-mtp-q4_K_M?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.