Model benchmarks
devstral-small-2:latest local LLM performance
As of April 2026, devstral-small-2:latest runs at up to 11.8 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
24.0B
Peak speed
11.8 tok/s
Average speed
11.1 tok/s
Min memory
25.1 GB
Max context
65.536 tokens
Avg quality
64.2
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 | 62.7 | 64.2 | 65.7 |
| Agent Workflow | 62.2 | 66.2 | 70.1 |
| Code Generation | 53.7 | 56.9 | 60.0 |
| Role Play & Narrative | 70.0 | 71.4 | 72.8 |
| Research & Analysis | 59.8 | 62.2 | 64.7 |
Performance by hardware and tool
Every hardware/tool/quantization combination devstral-small-2:latest has been benchmarked on, ranked by peak token generation speed. Last updated April 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XTX | Ollama | Q4_K_M | 11.8 tok/s | 11.1 tok/s | 25.1 GB | 65.536 tokens | 64.2 | 2 |
Benchmark runs
All 2 devstral-small-2:latest runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is devstral-small-2:latest good for coding?
- In our benchmarks, devstral-small-2:latest scores 56.9/100 for coding. It runs at about 11.1 tok/s, so if you want more speed, /home/polaris/models/Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL.gguf is faster (~35.6 tok/s) and still scores well for coding (84.9/100).
- Is devstral-small-2:latest good for agentic (tool-using) tasks?
- In our benchmarks, devstral-small-2:latest scores 66.2/100 for agentic workflows. It runs at about 11.1 tok/s, so if you want more speed, muse-glimmer:latest is faster (~34.5 tok/s) and still scores well for agentic workflows (93.0/100).
- How fast is devstral-small-2:latest for local inference?
- Across 2 community benchmark runs, devstral-small-2:latest reaches up to 11.8 tok/s and averages 11.1 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does devstral-small-2:latest need?
- The leanest observed configuration used about 25.1 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run devstral-small-2:latest?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.