Model benchmarks
Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M local LLM performance
As of November 2026, Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M runs at up to 18.2 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
3B
Peak speed
18.2 tok/s
Average speed
18.0 tok/s
Avg PP
1667.8 tok/s
Min memory
n/a
Max context
32,768 tokens
Avg output / run
5,644 tokens
Avg runtime / run
5m 43s
Avg quality
52.9
Benchmark runs
3
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 3 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 | 51.3 | 52.9 | 54.0 |
| Agent Workflow | 47.9 | 63.9 | 78.4 |
| Code Generation | 7.9 | 22.1 | 40.9 |
| Role Play & Narrative | 63.7 | 69.0 | 72.5 |
| Research & Analysis | 54.2 | 56.8 | 60.9 |
Performance by hardware and tool
Every hardware/tool/quantization combination Ministral-3-3B-Instruct-2512-GGUF-Q5_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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 2050 | Lemonade | Q5_K_M | 18.2 tok/s | 18.0 tok/s | n/a | 32,768 tokens | 52.9 | 3 |
Benchmark runs
All 3 Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M good for coding?
- In our benchmarks, Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M scores 22.1/100 for coding. It runs at about 18.0 tok/s, so if you want more speed, byteshape/Qwen3.8-27B-GGUF:IQ4_XS-ASCII-Condensed is faster (~42.0 tok/s) and still scores well for coding (86.3/100).
- Is Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M good for agentic (tool-using) tasks?
- In our benchmarks, Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M scores 63.9/100 for agentic workflows. It runs at about 18.0 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 Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M for local inference?
- Across 3 community benchmark runs, Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M reaches up to 18.2 tok/s and averages 18.0 tok/s, with the fastest results on NVIDIA GeForce RTX 2050.
- Which tools have been used to run Ministral-3-3B-Instruct-2512-GGUF-Q5_K_M?
- Benchmarks were submitted using Lemonade. Results are community-contributed and updated as new runs arrive.