Compare / LFM2.5 8B-A1B vs Apertus 1.5 8B
LFM2.5 8B-A1B vs Apertus 1.5 8B
Updated November 2026
Verdict
LFM2.5 8B-A1B scores higher on the Small LM Index: 40.1 vs 23.5. On the 1 benchmark both labs report, LFM2.5 8B-A1B wins 1 and Apertus 1.5 8B wins 0. Both need about 7.2 GB of memory at 4-bit quantization. Neither model has a hosted API on OpenRouter, so you would run both yourself.
Pick LFM2.5 8B-A1B if…
- you want the strongest model overall
Pick Apertus 1.5 8B if…
- you work with long documents
- you need a permissive license (Apache-2.0)
- you need image input
Specifications
| LFM2.5 8B-A1B | Apertus 1.5 8B | |
|---|---|---|
| Small LM Index | 40.1 (likely 37–43) | 23.5 (likely 16–34) |
| Lab | Liquid AI · United States | Swiss AI Initiative (Apertus) · Europe |
| Released | 2026-05-28 | 2026-07-24 |
| Parameters | 8.5B (1B active, MoE) | 8.9B (dense) |
| Memory at 4-bit | ~5.9 GB | ~7.2 GB |
| Context window | 128K tokens | 262K tokens |
| Input | text | text, image, audio |
| 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) | 31.3K | 22.8K |
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 8B-A1B | Apertus 1.5 8B |
|---|---|---|
| MMLU-Pro Knowledge | ≈ 63 | 48.8 |
| GPQA Diamond Knowledge | 51.3✓ | 27.2 |
| Humanity's Last Exam Reasoning | 6.9✓ | ≈ 4 |
| AIME Math | 42.5 | ≈ 12 |
| τ²-Bench Agentic | 16.1✓ | — |
| IFBench Instruction following | 55.6✓ | — |
| IFEval Instruction following | 91.8 | — |
| BFCL Agentic | 48.5 | — |
| MATH-500 Math | 88.8 | — |
| AA-LCR Long context | 0.0✓ | — |
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 8B-A1B or Apertus 1.5 8B?
LFM2.5 8B-A1B scores higher on the Small LM Index: 40.1 vs 23.5. On the 1 benchmark both labs report, LFM2.5 8B-A1B wins 1 and Apertus 1.5 8B wins 0.
Which is cheaper to run?
Neither model has a hosted API on OpenRouter, so you would run both yourself. Both need about 7.2 GB of memory at 4-bit quantization.
Which has the longer context window?
Apertus 1.5 8B has the longer context window: 262K tokens against 128K.
Can I use LFM2.5 8B-A1B and Apertus 1.5 8B commercially?
LFM2.5 8B-A1B is released under LFM Open License and Apertus 1.5 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.