Model benchmarks
unsloth/Qwen3.6-35B-A3B-GGUF local LLM performance
As of November 2026, unsloth/Qwen3.6-35B-A3B-GGUF runs at up to 48.7 tok/s for local inference (best of 2 community benchmark runs across 2 GPUs).
Model size
35B
Peak speed
48.7 tok/s
Average speed
46.4 tok/s
Avg PP
478.6 tok/s
Min memory
n/a
Max context
98,304 tokens
Avg output / run
14,450 tokens
Avg runtime / run
5m 24s
Avg quality
56.0
Benchmark runs
2
GPUs tested
2
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 | 37.2 | 56.0 | 74.8 |
| Agent Workflow | 18.0 | 47.7 | 77.3 |
| Code Generation | 43.7 | 55.2 | 66.7 |
| Role Play & Narrative | 75.8 | 78.7 | 81.6 |
| Research & Analysis | 5.6 | 42.4 | 79.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination unsloth/Qwen3.6-35B-A3B-GGUF 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3080 | Unsloth Studio | UD-Q4_K_XL | 48.7 tok/s | 48.7 tok/s | n/a | 98,304 tokens | 76.8 | 1 |
| NVIDIA GeForce RTX 4070 SUPER | llama.cpp | — | 44.2 tok/s | 44.2 tok/s | n/a | 65,536 tokens | 35.2 | 1 |
Benchmark runs
All 2 unsloth/Qwen3.6-35B-A3B-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is ggml-org/Qwen3.6-35B-A3B-GGUF good for coding?
- In our benchmarks, ggml-org/Qwen3.6-35B-A3B-GGUF scores 55.2/100 for coding. It runs at about 46.4 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 ggml-org/Qwen3.6-35B-A3B-GGUF good for agentic (tool-using) tasks?
- In our benchmarks, ggml-org/Qwen3.6-35B-A3B-GGUF scores 47.7/100 for agentic workflows. It runs at about 46.4 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 unsloth/Qwen3.6-35B-A3B-GGUF for local inference?
- Across 2 community benchmark runs, unsloth/Qwen3.6-35B-A3B-GGUF reaches up to 48.7 tok/s and averages 46.4 tok/s, with the fastest results on NVIDIA GeForce RTX 3080.
- Which tools have been used to run unsloth/Qwen3.6-35B-A3B-GGUF?
- Benchmarks were submitted using Unsloth Studio, llama.cpp. Results are community-contributed and updated as new runs arrive.