Compare / LFM2.5 350M vs Qwen3.5 0.8B
LFM2.5 350M vs Qwen3.5 0.8B
Updated November 2026
Verdict
LFM2.5 350M scores higher on the Small LM Index: 12.6 vs 10.7, although their likely ranges overlap, so the lead is not certain. On the 6 benchmarks both labs report, LFM2.5 350M wins 4 and Qwen3.5 0.8B wins 2. Both need about 1.2 GB of memory at 4-bit quantization. Neither model has a hosted API on OpenRouter, so you would run both yourself.
Pick LFM2.5 350M if…
- you want the strongest model overall
- you need precise instruction following
Pick Qwen3.5 0.8B if…
- you work with long documents
- you need a permissive license (Apache-2.0)
- you need image input
- you want the more widely used model (more quantized builds and community support)
Specifications
| LFM2.5 350M | Qwen3.5 0.8B | |
|---|---|---|
| Small LM Index | 12.6 (likely 10–17) | 10.7 (likely 9–13) |
| Lab | Liquid AI · United States | Alibaba (Qwen) · China |
| Released | 2026-03-31 | 2026-03-02 |
| Parameters | 350M (dense) | 900M (dense) |
| Memory at 4-bit | ~0.8 GB | ~1.2 GB |
| Context window | 32K tokens | 262K tokens |
| Input | text | text, image, video |
| Reasoning | No thinking | Hybrid (thinking on/off) |
| License | LFM Open License | Apache-2.0 |
| API price / 1M tokens | Self-host only | Self-host only |
| Hugging Face downloads (30 days) | 114K | 3.7M |
Benchmarks
Higher is better. Bold marks the winner where both labs report a score. ≈ marks the index model's estimate for a score that wasn't reported, ✓ an independent measurement and * a score measured by another lab.
| Benchmark | LFM2.5 350M | Qwen3.5 0.8B |
|---|---|---|
| MMLU-Pro Knowledge | 20.0 | 42.3 |
| GPQA Diamond Knowledge | 30.6 | 11.9 |
| τ²-Bench Agentic | 18.4 | 11.6 |
| IFBench Instruction following | 40.7 | 21.0 |
| MMMU-Pro Vision | — | 25.8✓ |
| MMLU Knowledge | — | 59.5 |
| Multilingual MMLU Multilingual | — | 44.3 |
| IFEval Instruction following | 77.0 | 44.0 |
| BFCL Agentic | 21.9 | 25.3 |
| AA-LCR Long context | — | 4.7 |
Scores are mostly reported by each lab with its own settings, so small gaps are noise. See the methodology. Open in the interactive comparison to add more models.
Questions
Which is better, LFM2.5 350M or Qwen3.5 0.8B?
LFM2.5 350M scores higher on the Small LM Index: 12.6 vs 10.7, although their likely ranges overlap, so the lead is not certain. On the 6 benchmarks both labs report, LFM2.5 350M wins 4 and Qwen3.5 0.8B wins 2.
Which is cheaper to run?
Neither model has a hosted API on OpenRouter, so you would run both yourself. Both need about 1.2 GB of memory at 4-bit quantization.
Which has the longer context window?
Qwen3.5 0.8B has the longer context window: 262K tokens against 32K.
Can I use LFM2.5 350M and Qwen3.5 0.8B commercially?
LFM2.5 350M is released under LFM Open License and Qwen3.5 0.8B under Apache-2.0. Apache 2.0 and MIT allow commercial use; other licenses can add conditions, so read the license before shipping a product.