Liquid AI ships LFM2.5-24B LoRA adapters for Pi and Claude Code, cutting tool calls 40%
yb2698 · x · 2026-10-31
Liquid AI released LoRA adapters trained on top of LFM2.5-24B-A2B, tuned for more efficient agentic behavior in open-source harnesses like Pi and Claude Code — up to 40% fewer tool calls on Pi. The company says LFMs are built for customization, arguing specialized intelligence covers more than half of enterprise AI use cases with predictable economics; the Hugging Face team collaborated on the tuning. Adapters are open-sourced on Hugging Face.
Related event: Liquid AI releases LFM2.5-24B harness adapters cutting tool calls up to 40%(7 posts)→
More from coding & agent
- Fighting app spam in the LLM era: risk scoring and checkpoints, says ex-Tinder exec — nikitabier · 2026-10-31
- Human-in-the-loop with Claude: rubber-stamping 400-line bash scripts as 'approval' — generativist · 2026-10-31
- How to actually evaluate a GraphRAG stack: multi-hop questions, conflicting facts, citations — JeremyCMorgan · 2026-10-31
- Suno ships MCP Beta, dev uses it to sync motion graphics to AI-generated music — enggirlfriend · 2026-10-31
- Count document passes, not tokens: batching questions cuts LLM scoring cost — petrusenko_max · 2026-10-31
- Open-source framework uses a second agent to catch AI slop before it hits your repo — kwerkee · 2026-10-31