LiquidAI's 2.6B model rivals 4x larger models on agentic tasks
LFM2.5 2.6B model competitive with 4x larger models
LiquidAI has released LFM2.5-2.6B, a compact 2.6B-parameter model that competes with models four times its size on agentic workloads like tool use and instruction following. It features a 128K context window, agentic reinforcement learning, and efficient inference, achieving 220 tokens/s on an Apple M5 Max and 113 tokens/s on an AMD Ryzen CPU, all in under 2.5 GB of memory. The model is part of the LFM2.5 family, designed for on-device deployment, and is available in multiple formats for various frameworks.
LFM2.5-2.6B is the fastest model we tested, with decode speeds of 220 tokens/s on an M5 Max and 113 tokens/s on a Ryzen AI Max+ 395.
- vezycash
Always put "Lower is better" or "higher is better" in benchmarks. Not everyone knows what your numbers mean.
- lend000
I can't imagine who is using something like this for agentic coding, but I see exciting opportunities on the horizon when we can have hundreds of reasonably rational and conversational agents working on local machines to simulate emergent behavior (simulating crowds, markets, ecosystems, game NPCs, etc.)
- Gecko4072
These LiquidAI models have never worked well for me in practice.
- lostmsu
It's not even competitive with 2x sized Qwen 4B.
Why is Qwen3.5 2B not in the table?
- 0xbadcafebee
LFM's training/post-training is famously different than other models. They target reliable operation of tiny models in ways other model families don't (they aren't just scaling a larger model to a smaller size). If you're looking for good performance out of tiny models, LFM has the most advanced design.
Note how they're much smaller than all other models in the comparison yet match or exceed them. This is for 2.6B params, but they have models as small as 230M. Nobody else designs models that small.