Compare / LFM2.5 2.6B vs Qwen3.5 2B
LFM2.5 2.6B vs Qwen3.5 2B
Updated November 2026
Verdict
LFM2.5 2.6B scores higher on the Small LM Index: 43.2 vs 33.0. On the 4 benchmarks both labs report, LFM2.5 2.6B wins 4 and Qwen3.5 2B wins 0. Both need about 2.3 GB of memory at 4-bit quantization. Only LFM2.5 2.6B has a hosted API on OpenRouter ($0.39 per million tokens); the other has to be self-hosted.
Pick LFM2.5 2.6B if…
- you want the strongest model overall
- you need precise instruction following
- you build agents that call tools
- you want a hosted API rather than running it yourself
Pick Qwen3.5 2B 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 2.6B | Qwen3.5 2B | |
|---|---|---|
| Small LM Index | 43.2 (likely 39–48) | 33.0 (likely 29–37) |
| Lab | Liquid AI · United States | Alibaba (Qwen) · China |
| Released | 2026-07-28 | 2026-03-02 |
| Parameters | 2.7B (dense) | 2.3B (dense) |
| Memory at 4-bit | ~2.3 GB | ~2 GB |
| Context window | 128K 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 | $0.39 ($0.35 in, $0.50 out) | Self-host only |
| Hugging Face downloads (30 days) | 69K | 3.4M |
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 2.6B | Qwen3.5 2B |
|---|---|---|
| MMLU-Pro Knowledge | ≈ 68 | 66.5 |
| GPQA Diamond Knowledge | 57.8 | 51.6 |
| Humanity's Last Exam Reasoning | 5.9 | ≈ 3 |
| AIME Math | 51.9 | ≈ 39 |
| LiveCodeBench Coding | 59.4 | ≈ 36 |
| SWE-bench Verified Coding | 18.0 | ≈ 14 |
| τ²-Bench Agentic | — | 48.8 |
| IFBench Instruction following | 59.2 | 41.3 |
| MMLU Knowledge | — | 79.6 |
| Multilingual MMLU Multilingual | — | 63.1 |
| IFEval Instruction following | 95.4 | 78.6 |
| BFCL Agentic | 56.9 | 43.6 |
| HMMT Math | — | 22.9 |
| AA-LCR Long context | — | 25.6 |
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 2.6B or Qwen3.5 2B?
LFM2.5 2.6B scores higher on the Small LM Index: 43.2 vs 33.0. On the 4 benchmarks both labs report, LFM2.5 2.6B wins 4 and Qwen3.5 2B wins 0.
Which is cheaper to run?
Only LFM2.5 2.6B has a hosted API on OpenRouter ($0.39 per million tokens); the other has to be self-hosted. Both need about 2.3 GB of memory at 4-bit quantization.
Which has the longer context window?
Qwen3.5 2B has the longer context window: 262K tokens against 128K.
Can I use LFM2.5 2.6B and Qwen3.5 2B commercially?
LFM2.5 2.6B is released under LFM Open License and Qwen3.5 2B 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.