LFM2.5-2.6B

· llama.cpp · 8K context · x86-64 CPU (AVX2) · System RAM

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

38 tok/sdecode

84 tok/sprefill

8Kcontext window

Tested on this cardOct 3, 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 LiquidAI/LFM2.5-2.6B-GGUF LFM2.5-2.6B-QAD-Q4_0.gguf \
  --revision 84022ce711b28455e8c4fc364ce68c00cf995875 \
  --local-dir ~/models/LFM2.5-2.6B-GGUF-84022ce7

2 Start the server

docker run --rm \
  -p 8000:8080 \
  --shm-size 1g \
  -v ~/models/LFM2.5-2.6B-GGUF-84022ce7:/models:ro \
  --entrypoint /app/llama-server \
  ghcr.io/ggml-org/llama.cpp@sha256:23fd59bc5e5b06ca68003a5772d441f0411fb723eeaff850b9111bcd2ac4fd33 \
  -m /models/LFM2.5-2.6B-QAD-Q4_0.gguf \
  --alias LFM2.5-2.6B-GGUF \
  --ctx-size 8192 \
  --parallel 1 \
  --n-gpu-layers 0 \
  --flash-attn off \
  --no-mmproj \
  --jinja \
  --reasoning-format deepseek \
  --threads 6 \
  --threads-batch 6 \
  --batch-size 512 \
  --ubatch-size 128 \
  --host 0.0.0.0 \
  --port 8080 \
  --metrics

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 AMD EPYC 7443P CPU, six threads, 4 GiB container cap, no GPU on Oct 3, 2026.

Details

Weights: LiquidAI/LFM2.5-2.6B-GGUF @ 84022ce711 ›
Image: ghcr.io/ggml-org/llama.cpp@sha256:23fd59bc5e5b06ca68003a5772d441f0411fb723eeaff850b9111bcd2ac4fd33
Launch: llama.cpp-lfm2.5-2.6b-qad-q4-0-8k-cpu-docker ›
Recipe file: registry/recipes/cpu/x86-64-avx2-cpu/lfm2.5-2.6b.llama.cpp.8k.json ›
Model released: Jul 28, 2026
Tested: owner, AMD EPYC 7443P CPU, six threads, 4 GiB container cap, no GPU