LFM2.5-230M-Base
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
230 M
Context
128,000
Downloads 30d
149 K
Likes
24
Download history
tracking started — chart appears after 7 days of snapshots (3 recorded)149 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 | 0.5 GB | 1.0 GB | ✅ Runs comfortably |
| model.safetensors (bf16, full) | bf16 + 125K ctx | 0.5 GB | 1.5 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/lfm2-5-230m-base
{
"hf_id": "LiquidAI/LFM2.5-230M-Base",
"params_b": 0.23,
"context_length": 128000,
"license": { "id": "other", "commercial": "unknown" },
"downloads_30d": 148905,
"vram_estimates": [
{ "quant": "bf16", "gb": 1.0 }
],
"updated_at": "2026-10-23T01:00:19Z"
}
Specifications
- Architecture
- Lfm2ForCausalLM
- Parameters
- 230 M
- Tensor type
- BF16
- Context length
- 128,000
- Vocabulary
- 65,536
- Layers / heads
- 14 / 16
- Licence
- other
- First seen on the Hub
- 2026-06-16
- Training datasets
- undisclosed
- Added to our catalog
- 2026-10-23
Compare with any text-generation model