# Liquid AI releases 24-billion-parameter LFM2.5 model

> The company says two billion parameters are active per token and reports device-specific results.

- Canonical URL: https://newsverum.com/en/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026
- Publisher: Newsverum
- Section: Technology
- Country: United States
- Published: 2026-10-15T18:13:44+00:00
- Updated: 2026-10-15T18:13:44+00:00
- Event / data period: 2026-10-13
- Other editions: [Türkçe](https://newsverum.com/tr/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md), [العربية](https://newsverum.com/ar/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md), [Français](https://newsverum.com/fr/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md), [Español](https://newsverum.com/es/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md), [Deutsch](https://newsverum.com/de/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md), [Português](https://newsverum.com/pt/news/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026.md)

Liquid AI released LFM2.5-24B-A2B on October 13 as an open-weight reasoning model for local agents. The mixture-of-experts model has 24 billion total parameters and activates two billion per token. Liquid AI says its context window grew from 32,768 to 131,072 tokens, vocabulary from 65,000 to 128,000, and pretraining from 17 trillion to 33 trillion tokens versus its predecessor.

The company reports 184 generated tokens per second on an Apple M5 Max and more than 21,000 output tokens per second on one NVIDIA B300. The latter is a high-concurrency sustained-load result, not a single-user rate; both figures come from Liquid AI’s benchmarking setup. Its benchmark table shows mixed results, with the model ahead or tied on five of eight tool-use tests but behind larger models on several knowledge and coding tests.

## Data

### Announcement summary

The company says two billion parameters are active per token and reports device-specific results.

Methodology: This event record summarizes the source announcement and adds no independent verification or generalization beyond it.

| date | label | description | status |
|---|---|---|---|
| 2026-10-13 | Liquid AI releases 24-billion-parameter LFM2.5 model | The company says two billion parameters are active per token and reports device-specific results. | occurred |

Sources: <https://www.liquid.ai/blog/lfm2-5-24b-a2b> · Download data: [JSON](https://newsverum.com/api/v1/articles/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026/data/announcement-record?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/liquid-ai-releases-lfm25-24b-a2b-on-device-model-2026/data/announcement-record?lang=en&format=csv)

## Sources

- [Liquid AI](<https://www.liquid.ai/blog/lfm2-5-24b-a2b>) (Source checked: 2026-10-15; Published: 2026-10-13)

---

Cite the canonical URL above. [Editorial standards and corrections](https://newsverum.com/en/editorial)