Get the latest setups
Models/LFM2.5 VL 3B DSpark/LFM2.5 VL 3B DSpark GGUF
Text / LiquidAI

LFM2.5 VL 3B DSpark GGUF

LFM2.5 VL 3B DSpark GGUF is a GGUF build of LiquidAI/LFM2.5-VL-3B-DSpark for llama.cpp. It is a standalone draft sidecar containing only a drafter with 4 attention layers, a Markov head, a confidence head, and block size 9, used for speculative decoding with a compatible LFM2.5-VL-3B-GGUF target model. Token embeddings and the LM head are shared from the target model at load time, so pairing with the target file is required. The supplied F16 drafter file is 567 MB. Speculative decoding is described as exact under greedy decoding because the target verifies every proposed token.

EstimatedRepository opened Sep 18, 2026Source checked 9/28/2026Version: 5ff73699
LOCALRENTED GPUOPEN WEIGHTS
Freshness 90 out of 100. Repository age: 27 days old, Sep 18, 2026. Measures age, not quality.
Popularity 41 out of 100. 1,608 downloads in 30 days; 12 lifetime likes. Checked Oct 15, 2026. Measures adoption, not quality.
What do these scores mean?

Two ways to compare models. Neither score measures answer quality, benchmark performance, or whether a model fits your machine.

Freshness · release age

A new release starts at 100. The score halves every 180 days: 100 × 2^(−age in days / 180). When the release date is unavailable, a repository date is a fallback—not proof of when the model was released. Missing dates stay unscored.

Check the date source
Popularity · adoption

Downloads over the past 30 days carry 80% of the score; lifetime likes carry 20%. Both use a logarithmic scale, capped at 10 million downloads and 100,000 likes. If likes are unavailable, downloads carry all the weight. Missing download statistics stay unscored.

round(80 × min(1, log10(1 + downloads) / 7) + 20 × min(1, log10(1 + likes) / 5))

Check the statistics source
Scores as of Oct 15, 2026. Statistics checked Oct 15, 2026. Statistics are marked for refresh after 72 hours.Full scoring method

At a glance

Architecture
unknown
License
lfm1.0
Software
llama.cpp
View the model source
MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

  • Running speculative decoding with a paired LFM2.5-VL-3B-GGUF target model in llama.cpp.
  • Serving image-text-to-text tasks with the target plus drafter setup in llama.cpp.
  • Using the drafter as a sidecar for the LFM2.5-VL-3B-GGUF target model.
  • Testing draft acceptance rate through llama.cpp timing logs when paired with a compatible target.
  • Using the GGUF drafter on Apple silicon with recommended block size 8 via --spec-draft-n-max 8.

Tradeoffs

  • Not a standalone model; it must be paired with the LFM2.5-VL-3B-GGUF target file, which supplies token embeddings and the LM head.
  • Requires a compatible LFM2.5-VL-3B-GGUF target; other target models are not stated as supported.
  • It is a draft sidecar for speculative decoding, not a full vision-language model by itself.
  • The supplied file is an F16 drafter only; no other quantization files are listed.
  • Performance, compatibility, hardware requirements beyond the provided run command, and purchase suitability are not stated in the supplied facts.

Ways to run it

llama.cpp · Windows, macOS, Linux

Repository-specific command in publisher documentation; confirm dependencies and hardware in the source.

llama-server -hf LiquidAI/LFM2.5-VL-3B-GGUF:F16 -hfd LiquidAI/LFM2.5-VL-3B-DSpark-GGUF:F16 --spec-type draft-dspark --spec-draft-n-max 8 --spec-draft-n-min 0 -ngl 99 -ngld 99 -fa on
Official instructions

Explore its uses

GGUFIMAGE-TEXT-TO-TEXT

Keep exploring

Text
Prism ML

Bonsai 2 · 27B

A compressed 27B-class reasoning model with publisher-provided GGUF packs and a dedicated llama.cpp fork for CUDA, Metal, and CPU.

Freshness 90 out of 100. Release age: 28 days old, Sep 17, 2026. Measures age, not quality.
Popularity 90 out of 100. 4,849,199 downloads in 30 days; 2,697 lifetime likes. Checked Oct 15, 2026. Measures adoption, not quality.
MODEL SIZE27.36B
LOCALRENTED GPU
Released Sep 17, 2026
Text
DeepSeek

DeepSeek V4.1 Flash

A multimodal reasoning model with a compressed key-value cache, published as open weights.

Freshness 87 out of 100. Repository age: 35 days old, Sep 10, 2026. Measures age, not quality.
Popularity 84 out of 100. 1,238,583 downloads in 30 days; 4,519 lifetime likes. Checked Oct 15, 2026. Measures adoption, not quality.
Arena Open Source Text: rank 5 of 225. Ranked model: deepseek-v4.1-flash-max. Export dated Oct 13, 2026. Source checked Oct 15, 2026. Human preference ranking, not hardware suitability.+3
LOCALRENTED GPU
Repository opened Sep 10, 2026
Text
Qwen

Qwen3.8

Explore 4 versions of Qwen3.8. Compare model sizes, packages, hardware requirements, and available setups.

4 versionsCompare sizes & formats
LOCALRENTED GPU
Compare versions →
LFM2.5 VL 3B DSpark GGUF: hardware, VRAM & setup | YouRunAI