PulseAugur
EN
LIVE 18:35:20

Liquid AI's LFM2.5-24B-A2B model offers fast inference but high RAM needs

By PulseAugur Editorial · [1 sources] ·

Liquid AI has released its LFM2.5-24B-A2B model, a 24 billion parameter mixture-of-experts model that only utilizes 2.45 billion parameters per token. This architecture allows for faster decoding speeds, with Liquid reporting 184 tokens/s on an Apple M5 Max. Despite the active parameter efficiency, the full model still requires approximately 14.5 GB of RAM for storage. The model excels in instruction following and tool-calling tasks, outperforming competitors like Qwen3.6-35B-A3B in these areas, though it lags in knowledge-based benchmarks. AI

IMPACT This model's efficient parameter usage could influence future LLM architectures for edge devices, though high RAM requirements remain a constraint.

RANK_REASON The article details the local execution and performance characteristics of a specific mixture-of-experts LLM, focusing on practical usage and hardware requirements.

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Liquid AI's LFM2.5-24B-A2B model offers fast inference but high RAM needs

How we ranked this

Signal score
30 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The article details the local execution and performance characteristics of a specific mixture-of-experts LLM, focusing on practical usage and hardware requirements.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Alan West ·

    How to Run LFM2.5-24B-A2B Locally: Only 2B Active, Still Needs 14.5 GB of RAM

    <p>LFM2.5-24B-A2B is Liquid AI's new 24B mixture-of-experts model, and it touches only 2.45B of its parameters per token. Liquid says it decodes at 184 tokens/s on an Apple M5 Max. You can run the GGUF locally in Ollama or llama.cpp with one command. It came out on October 13 and…

RELATED ENTITIES

RELATED TOPICS