LFM2.5-24B-A2B

Liquid AI · released Oct 14, 2026

Unknown

Openness taxonomy

access, licence, and disclosure — never a single open/closed binary

Weights availability: Described as open-weight
Licence: unknown
Licence category: unknown
Commercial use allowed: unknown
Derivatives allowed: unknown
Redistribution allowed: unknown
Field-of-use restrictions: unknown
Training code availability: unknown
Inference code availability: unknown
Architecture disclosure: unknown
Training data disclosure: unknown

openness confidence

The source describes the model as open-weight.

Deployability

raw values as extracted — free text, never inferred from branding

Self-hosting: unknown
On-premise: unknown
Edge deployability: Targets on-device use
API availability: unknown
Regions available: unknown

Benchmarks

as reported — vendor-supplied unless otherwise noted

LiveClaw 48.09Decoding speed on Apple M5 Max 184 tokens/s

National-security pathway & evidence

from the assessed source item, when one is linked

Impact: lowConfidence: lowNat-sec: noneindependent reporting

Evidence limitations: Single press report; the model's performance figures and open-weight status are not independently verified here, and no licence terms are provided.

Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active MoE for On-Device Agents

Linked research papers

from paper_model_links

No linked research papers yet.