SmolLM3: pricing, context window and benchmarks
SmolLM3 is a open-weight model from Hugging Face. It offers a 8,192-token context window, costs $1.71 per million input tokens and $8.86 per million output tokens, and averages 63.1 across the index's benchmark suite.
Pricing
SmolLM3 costs $1.71 per million input tokens and $8.86 per million output tokens. A blended 3:1 input-to-output mix works out to about $3.50 per million tokens. Cached input is listed at $0.23 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $1.71 | per 1M tokens |
| Output | $8.86 | per 1M tokens |
| Cached input | $0.23 | per 1M tokens |
| Blended 3:1 | $3.50 | estimate per 1M tokens |
| 1M in + 1M out | $10.57 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 69.2 | 405 | 33th | |
| GPQA Diamond | 61.0 | 398 | 34th | |
| SWE-bench Verified | 47.3 | 444 | 26th | |
| HumanEval | 69.5 | 403 | 33th | |
| MATH-500 | 68.1 | 400 | 33th | |
| AIME 2025 | 52.9 | 436 | 27th | |
| LiveCodeBench | 49.8 | 451 | 25th | |
| IFEval | 72.7 | 421 | 30th | |
| TruthfulQA | 67.9 | 415 | 31th | |
| HellaSwag | 77.7 | 405 | 33th | |
| MMMU | 58.3 | 454 | 24th | |
| Arena Elo | 1294.1 | 427 | 29th |
What SmolLM3 is good for
SmolLM3 is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $1.71 per million input tokens (hosted list price) and a 8,192-token context window, it is a practical option for teams that want more control than a closed API allows.
Strengths
- Input price of $1.71 per million tokens is at or below the index median.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
Trade-offs
- Context window of 8,192 tokens is small for long-document work.
- Average benchmark score is below the index median; test it on your own task.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Fusion Mini | OpenRouter | 4096 | $4.89 | 63.3 |
| DeepSeek-V3.2 7B | DeepSeek | 4096 | $1.75 | 63.2 |
| Firefunction Mini | Fireworks AI | 32768 | $4.63 | 63.2 |
| GLM-4.6 8B | Zhipu AI | 8192 | $1.69 | 63.1 |
| SmolLM3 | Hugging Face | 8192 | $1.71 | 63.1 |
| DeepSeek-R1 7B | DeepSeek | 32768 | $1.79 | 63.0 |
FAQ
How much does SmolLM3 cost?
SmolLM3 is listed at $1.71 per million input tokens and $8.86 per million output tokens. Cached input is $0.23 per million tokens. A blended 3:1 input-to-output mix is about $3.50 per million tokens.
What is the context window of SmolLM3?
SmolLM3 has a context window of 8,192 tokens and can emit up to 8,192 output tokens per request in this index.
Is SmolLM3 open weights?
Yes. SmolLM3 is listed as an open-weight model under a open weights licence.
How does SmolLM3 score on benchmarks?
SmolLM3 averages 63.1 across the scored benchmarks in this index, ranking #419 of 600 models. Its strongest relative result is Arena Elo at 1294.1.
What is SmolLM3 best for?
SmolLM3 is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $1.71 per million input tokens (hosted list price) and a 8,192-token context window, it is a practical option for teams that want more control than a closed API allows.