Model benchmarks
smollm2-1.7b-instruct local LLM performance
As of October 2026, smollm2-1.7b-instruct runs at up to 50.3 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
1.7B
Peak speed
50.3 tok/s
Average speed
49.4 tok/s
Avg PP
1955.8 tok/s
Min memory
3.7 GB
Max context
32,768 tokens
Avg output / run
3,594 tokens
Avg runtime / run
1m 20s
Avg quality
35.8
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 | 32.3 | 35.8 | 41.3 |
| Agent Workflow | 35.6 | 40.9 | 45.0 |
| Code Generation | 7.5 | 19.8 | 39.2 |
| Role Play & Narrative | 45.1 | 49.2 | 54.0 |
| Research & Analysis | 29.1 | 33.3 | 38.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination smollm2-1.7b-instruct 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 2050 | LM Studio | Q8_0 | 50.3 tok/s | 49.4 tok/s | 3.7 GB | 32,768 tokens | 35.8 | 3 |
Benchmark runs
All 3 smollm2-1.7b-instruct runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is smollm2-1.7b-instruct good for coding?
- In our benchmarks, smollm2-1.7b-instruct scores 19.8/100 for coding. It runs at about 49.4 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 smollm2-1.7b-instruct good for agentic (tool-using) tasks?
- In our benchmarks, smollm2-1.7b-instruct scores 40.9/100 for agentic workflows. It runs at about 49.4 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 smollm2-1.7b-instruct for local inference?
- Across 3 community benchmark runs, smollm2-1.7b-instruct reaches up to 50.3 tok/s and averages 49.4 tok/s, with the fastest results on NVIDIA GeForce RTX 2050.
- How much memory does smollm2-1.7b-instruct need?
- The leanest observed configuration used about 3.7 GB of memory (quantizations tested: Q8_0).
- Which tools have been used to run smollm2-1.7b-instruct?
- Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.