Today, we’re releasing LFM2.5-8B-A1B, a high-throughput edge model optimized for fast, reliable tool calling and complex instruction following on consumer hardware, delivering compressed performance competitive with much larger models and day-one support across major inference frameworks.
Sounds too good to be true, will probably test it though. The Gemma 4 E4B performs quite well, both in performance and in quality. Not good enough for a lot of cases though
Sounds too good to be true, will probably test it though. The Gemma 4 E4B performs quite well, both in performance and in quality. Not good enough for a lot of cases though