Model benchmarks
MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 local LLM performance
As of October 2026, MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 runs at up to 33.1 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
9B
Peak speed
33.1 tok/s
Average speed
32.2 tok/s
Avg PP
204.2 tok/s
Min memory
7.0 GB
Max context
65,536 tokens
Avg output / run
12,443 tokens
Avg runtime / run
7m 55s
Avg quality
63.9
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 | 62.6 | 63.9 | 65.2 |
| Agent Workflow | 66.7 | 73.0 | 79.2 |
| Code Generation | 48.6 | 48.8 | 48.9 |
| Role Play & Narrative | 62.7 | 63.8 | 64.9 |
| Research & Analysis | 68.3 | 70.1 | 71.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Apple M1 Pro | oMLX | FP16 | 33.1 tok/s | 32.2 tok/s | 7.0 GB | 65,536 tokens | 63.9 | 2 |
Benchmark runs
All 2 MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 good for coding?
- In our benchmarks, MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 scores 48.8/100 for coding. It runs at about 32.2 tok/s, so if you want more speed, Qwen3.8-27B-Turbo-Brilliance-50X-RI-GSQ-RCO-IQ3_XXS-mtp is faster (~37.5 tok/s) and still scores well for coding (83.8/100).
- Is MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 good for agentic (tool-using) tasks?
- In our benchmarks, MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 scores 73.0/100 for agentic workflows. It runs at about 32.2 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
- How fast is MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 for local inference?
- Across 2 community benchmark runs, MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 reaches up to 33.1 tok/s and averages 32.2 tok/s, with the fastest results on Apple M1 Pro.
- How much memory does MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16 need?
- The leanest observed configuration used about 7.0 GB of memory (quantizations tested: FP16).
- Which tools have been used to run MiMo-V2.6-Distill-Qwen-9B-oQ4e-fp16?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.