Model benchmarks
1bit-MONSTER/SmolLM3-3B-GGUF local LLM performance
As of November 2026, 1bit-MONSTER/SmolLM3-3B-GGUF runs at up to 48.9 tok/s for local inference (best of 6 community benchmark runs across 3 GPUs).
Model size
3B
Peak speed
48.9 tok/s
Average speed
34.3 tok/s
Avg PP
964.3 tok/s
Min memory
3.3 GB
Max context
65,536 tokens
Avg output / run
23,176 tokens
Avg runtime / run
15m 2s
Avg quality
31.9
Benchmark runs
6
GPUs tested
3
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 6 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 | 12.1 | 31.9 | 48.2 |
| Agent Workflow | 20.1 | 49.0 | 79.1 |
| Code Generation | 0.4 | 9.3 | 15.0 |
| Role Play & Narrative | 7.0 | 41.0 | 67.9 |
| Research & Analysis | 11.9 | 28.2 | 39.8 |
Performance by hardware and tool
Every hardware/tool/quantization combination 1bit-MONSTER/SmolLM3-3B-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 2050 | Lemonade | — | 48.9 tok/s | 48.8 tok/s | n/a | 16,384 tokens | 41.1 | 2 |
| Intel(R) Arc(TM) 140V GPU (16GB) | Unsloth Studio | UD-IQ2_XXS | 47.6 tok/s | 47.6 tok/s | 3.3 GB | 8,192 tokens | 6.3 | 1 |
| Intel(R) Arc(TM) 140V GPU (16GB) | Unsloth Studio | UD-Q4_K_XL | 30.1 tok/s | 30.1 tok/s | 9.6 GB | 65,536 tokens | 29.3 | 1 |
| AMD Radeon Vega Series / Radeon Vega Mobile Series | Unsloth Studio | Q4_K_M | 15.2 tok/s | 15.1 tok/s | 4.5 GB | 65,536 tokens | 36.8 | 2 |
Benchmark runs
All 6 1bit-MONSTER/SmolLM3-3B-GGUF runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is unsloth/SmolLM3-3B-GGUF good for coding?
- In our benchmarks, unsloth/SmolLM3-3B-GGUF scores 9.3/100 for coding. It runs at about 34.3 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 unsloth/SmolLM3-3B-GGUF good for agentic (tool-using) tasks?
- In our benchmarks, unsloth/SmolLM3-3B-GGUF scores 49.0/100 for agentic workflows. It runs at about 34.3 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 1bit-MONSTER/SmolLM3-3B-GGUF for local inference?
- Across 6 community benchmark runs, 1bit-MONSTER/SmolLM3-3B-GGUF reaches up to 48.9 tok/s and averages 34.3 tok/s, with the fastest results on NVIDIA GeForce RTX 2050.
- How much memory does 1bit-MONSTER/SmolLM3-3B-GGUF need?
- The leanest observed configuration used about 3.3 GB of memory (quantizations tested: Q4_K_M, UD-IQ2_XXS, UD-Q4_K_XL).
- Which tools have been used to run 1bit-MONSTER/SmolLM3-3B-GGUF?
- Benchmarks were submitted using Lemonade, Unsloth Studio. Results are community-contributed and updated as new runs arrive.