LFM2.5-24B-A2B
4stories this week
5last 30 days
5all time
In the model registry
| Model | Params | Context | Released |
|---|---|---|---|
| LiquidAI/LFM2.5-24B-A2B | 24B | 128K | Oct 12, 2026 |
| LiquidAI/LFM2.5-24B-A2B-GGUF | 24B | – | Oct 9, 2026 |
Timeline
- Oct 19, 2026 · Model release · 2 sourcesNemotron 3.5 Bolt, Muse Spark Open, and a Fast LFM2.5 24BNemotron 3.5, with a closer look at Bolt and its “native 2-bit” quantization
- Oct 18, 2026 · Opinion / analysis · 1 sourceNine models, five coding tasks, one 16GB GPUI ran five coding tasks against nine models on one RX 9060 XT (16 GB).
- Oct 15, 2026 · Opinion / analysis · 1 sourceVery impressed with LFM2.5-24-A2B in Pi AgentJust downloaded the new LFm2.5-24B-A2B LiquidAI model for testing.
- Oct 14, 2026 · Model release · 3 sourcesRun fast, private agents on-device with LFM2.5-24B-A2BBest-in-class agent: Competitive with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks, scoring 48.09 on LiveClaw versus 38.19.
- Oct 12, 2026 · Model release · 1 sourceLiquidAI/LFM2.5-24B-A2BLiquidAI published the model LFM2.5-24B-A2B on Hugging Face.