LFM2.5-350M
LFM2.5 is a new family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with extended pre-training and reinforcement learning.
Params
355 M
Context
128,000
Downloads 30d
161 K
Likes
443
Download history
tracking started — chart appears after 7 days of snapshots (5 recorded)161 K downloads in the last 30 days
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | bf16 | 0.7 GB | 1.3 GB | ✅ Runs comfortably |
| model.safetensors (bf16, full) | bf16 + 125K ctx | 0.7 GB | 2.1 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ curl -s https://aimodelscomparison.com/api/v1/models/lfm2-5-350m
{
"hf_id": "LiquidAI/LFM2.5-350M",
"params_b": 0.35,
"context_length": 128000,
"license": { "id": "other", "commercial": "unknown" },
"downloads_30d": 161051,
"vram_estimates": [
{ "quant": "bf16", "gb": 1.3 }
],
"updated_at": "2026-11-06T01:00:11Z"
}
Specifications
- Architecture
- Lfm2ForCausalLM
- Parameters
- 355 M
- Tensor type
- BF16
- Context length
- 128,000
- Vocabulary
- 65,536
- Layers / heads
- 16 / 16
- Licence
- other
- First seen on the Hub
- 2026-03-31
- Training datasets
- undisclosed
- Added to our catalog
- 2026-11-06
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model