LFM2.5-24B-A2B
Open-weight sparse model from Liquid AI for local agents; day-one llama.cpp (GGUF), MLX, vLLM, SGLang and ONNX support.
Price$0.10 in, $1 out per 1M tokens
Checked15 October 2026
No published training cutoff
At least37events since its release on 13 October 2026
0 historic, 7 major. 2 days of news.
Liquid AI doesn’t publish a training cutoff, so we count from its release date. The real gap is larger.
How to call it
3 ways to use LFM2.5-24B-A2B.
| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| Hugging Face | huggingface.co | ||
| Liquid AI Console | console.liquid.ai | ||
| Liquid Playground | playground.liquid.ai |
Pricing
| Input | $0.10 per 1M tokens |
|---|---|
| Output | $1 per 1M tokens |
As published: “per 1M tokens”.
Source: liquid.ai, checked 15 October 2026.
Specs
- Input
- Text
- Output
- Text
- Context window
- 131,072 tokens
- Max output
- Not published
- Open weights
- Yes
- Licence
- lfm1.0
- Training cutoff
- Not published
- Released
- 13 October 2026
- Status
- Current
- Type
- Reasoning model
What stands out
What its maker marketed at launch, and what people found later, each with its source.
On-device agent MoE
24B total / 2B active parameters; 184 tokens/s on an Apple M5 Max and vendor-reported AutomationBench 27.69 vs 8.72 for Qwen3.6-35B-A3B.
Notes
Reasoning-only; cached input $0.05 per 1M. Weaker than Qwen3.6-35B-A3B on GPQA Diamond and SWE-bench Verified (vendor numbers).
Not yet checked against an official source: Console model id not verified.