Liquid AI's LFM2.5-VL-3B Beats Models Twice Its Size Running on Your Phone
Liquid AI's 3B vision-language model beats models up to 2.6x its size on screen understanding and grounding, running on-device at 20 tokens/s on a phone.
- Liquid AI releases LFM2.5-VL-3B, a 3.1B vision-language model that beats models up to 2.6x its size on screen understanding and grounding.
- Runs on-device: 228 tokens/s on Apple M5 Max, 20 tokens/s on a Galaxy S26 Ultra, in ~3 GB of memory.
- Scores 80.7 on ScreenSpot-v2 (UI grounding), ahead of Gemma-4-E4B at 51.2 and Qwen 3.5 4B at 78.5.
- Built on a hybrid convolution-attention backbone (not a pure transformer), cutting KV cache by up to 90% for edge efficiency.
- Ships with day-one support for llama.cpp, MLX, vLLM, SGLang, and ONNX; open weights on Hugging Face.
- Reaches ~11K output tokens/s at high concurrency on a single H100, roughly 2x faster than 4B-class competitors.
Liquid AI just released LFM2.5-VL-3B, a 3.1B-parameter vision-language model (VLM) designed to run fast on real hardware, from a single H100 GPU down to a Galaxy S26 Ultra. The headline claim is striking: it matches or beats models up to 2.6x its size on key benchmarks, while fitting in about 3 GB of memory and decoding at 228 tokens per second on an Apple M5 Max.
This is not a reasoning model. It answers directly, which keeps latency low. That design choice is deliberate: the target is real-time and on-device applications where you cannot afford a 10-second thinking trace before getting a result.
A non-transformer running on your phone
To understand why LFM2.5-VL-3B is fast, you need to understand what it is built on. The core of LFM2.5 is a hybrid of attention and short-range convolutions, where attention blocks account for only about 37% of the model, with the remaining 63% made up of cheaper convolution blocks. This matters because standard transformers hit a memory wall on edge devices because their KV cache grows with every token, while LFM2's architecture cuts that cache by up to 90% by replacing most attention layers with zero-cache convolution blocks.
Liquid AI is an AI startup spun out of MIT's CSAIL, founded in 2022, building foundation models for edge devices based on research into liquid neural networks. The VL model pairs this hybrid language backbone with a vision encoder: a SigLIP2 NaFlex encoder to convert input images into token sequences, using the 400M shape-optimized variant for fine-grained vision capabilities. NaFlex (Native Flex) is Google's image encoder that supports multiple resolutions and preserves the native image aspect ratio, which has the potential to improve aspect-sensitive applications such as document understanding.
The connector between vision and language is a lightweight 2-layer MLP that applies a PixelUnshuffle operation, projecting patch-level embeddings into language space by first lowering the number of visual tokens, then mapping image embeddings into the LFM2 hidden dimension. This is what keeps the token budget small and first-token latency low.
Four things that actually got better
LFM2.5-VL-3B builds on the previous LFM2-VL-3B with significant improvements in screen understanding, grounding, function calling, and multi-image input. Each of these is a meaningful jump, not a marginal one:
- Screen and UI understanding: It averages 80.7 on ScreenSpot-v2, far ahead of the much larger Gemma-4-E4B (51.2) and Qwen 3.5 4B (78.5) and close behind the larger InternVL-3.5-4B (84.1).