LFM2.5-2.6B

· vllm · 8K context · GTX 1650 SUPER · 4 GB

Liquid AI's 2.6B hybrid model: tiny and fast, built for on-device work.

20 tok/sdecode

53 tok/sprefill

8Kcontext window

Tested on this cardOct 23, 2026

Run it

Follow the steps below to download and run this recipe. Omarchy Local AI lists the recipes it can launch on your hardware. The server then answers on http://localhost:8000/v1.

1 Download the weights

hf download plavno/LFM2.5-2.6B-AutoRound-W4A16 \
  --revision 2244dd328189f0b58d9b938b576937ad7d216b7f \
  --local-dir ~/models/LFM2.5-2.6B-AutoRound-W4A16-2244dd32

2 Start the server

docker run --rm \
  --gpus all \
  -p 8000:8000 \
  --shm-size 16g \
  -v ~/models/LFM2.5-2.6B-AutoRound-W4A16-2244dd32:/models/LFM2.5-2.6B-AutoRound-W4A16-2244dd32:ro \
  --entrypoint vllm \
  vllm/vllm-openai:v0.30.0@sha256:8a69ffad015f138d7170c4ddc429e230a3bc1c1719f67e14324749df200a4b90 \
  serve \
  /models/LFM2.5-2.6B-AutoRound-W4A16-2244dd32 \
  --served-model-name LFM2.5-2.6B-AutoRound-W4A16-2244dd32 \
  --host 0.0.0.0 \
  --port 8000 \
  --max-model-len 8192 \
  --max-num-seqs 1 \
  --gpu-memory-utilization 0.78 \
  --kv-cache-dtype auto \
  --enable-auto-tool-choice \
  --tool-call-parser lfm2 \
  --reasoning-parser deepseek_r1

What it passed

✓ loadLoads and serves within an hour

✓ chatAnswers a plain question and stops on its own

✓ reasoningThinks separately and gets 17 × 23 right

✓ toolsCalls a tool with the right arguments and uses the result

✓ contextRecalls a code buried in a prompt that fills 85% of the window

✓ speedDecodes at 15 tok/s or more

Passed all six checks on a real NVIDIA GeForce GTX 1650 SUPER on Oct 23, 2026.

Details

Weights: plavno/LFM2.5-2.6B-AutoRound-W4A16 @ 2244dd3281 ›
Image: vllm/vllm-openai:v0.30.0@sha256:8a69ffad015f138d7170c4ddc429e230a3bc1c1719f67e14324749df200a4b90
Engine profile: vllm ›
Recipe file: registry/recipes/nvidia/gtx-1650-super-4gb/lfm2.5-2.6b.vllm.8k.json ›
Model released: Jul 28, 2026
Tested: owner, NVIDIA GeForce GTX 1650 SUPER