SmolLM2 1.7B
The small-model standout: 1.7B, Apache-2.0, runs anywhere.
About
SmolLM2-1.7B-Instruct is Hugging Face's tiny-but-mighty model: 1.7B parameters, Apache-2.0 licensed, trained on 11T tokens with strong instruction-following for its size — plus ONNX variants for phones, browsers (transformers.js), and edge devices. 223k HF downloads, not gated. It proved small models could be genuinely useful assistants, powering on-device chat, RAG-lite, and function calling where a 7B won't fit. Capacity limits reasoning depth; SmolLM3 now leads the family. But for "runs anywhere" Apache-2.0 intelligence, this remains the reference.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported (HF, Oct 2024): strong 1-2B-class on MMLU, ARC, HellaSwag; function-calling support — provider-reported.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: HF downloads 223,260 (peer-normalized); released 712 days ago. Manual review: small-model standout (1.7B, ONNX variants, runs on phones), 223k dl; 1.7B limits reasoning; superseded by SmolLM3. License: Apache-2.0.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.9/10
- Maintenance · recency, releases, issue load
- 5.0/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · HF downloads 223,260 (peer-normalized)
- · released 712 days ago
- · manual review: small-model standout (1.7B, ONNX variants, runs on phones), 223k dl; 1.7B limits reasoning; superseded by SmolLM3