Get the latest setups
Models/LFM2.5 1.2B Instruct DSpark
Text / LiquidAI

LFM2.5 1.2B Instruct DSpark

LFM2.5 1.2B Instruct DSpark is a speculative-decoding draft model for LiquidAI/LFM2.5-1.2B-Instruct, using the DSpark method adapted to the LFM2.5 architecture. It has 295.7M draft parameters in BF16, a 5-layer full-attention backbone, a Markov head (rank 256), and a confidence head, with a block size of 9 and vocabulary of 65,536. The model is intended to accelerate text generation with the target LFM2.5-1.2B-Instruct model in SGLang and is also described as running on Apple silicon via the Metal backend. The official model card reports measured speedups and acceptance rates on H100 and M4 Max hardware.

EstimatedRepository opened Aug 10, 2026Source checked 9/24/2026Version: 4876d048
LOCALRENTED GPUOPEN WEIGHTS
Freshness 70 out of 100. Repository age: 91 days old, Aug 10, 2026. Measures age, not quality.
Popularity 43 out of 100. 1,735 downloads in 30 days; 35 lifetime likes. Checked Nov 8, 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 Nov 9, 2026. Statistics checked Nov 8, 2026. Statistics are marked for refresh after 72 hours.Full scoring method

At a glance

Parameters
0.30B
Architecture
qwen3
Context length
128,000
License
lfm1.0
Disk space
0.6 GB
Software
SGLang
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

  • Accelerating text generation with LiquidAI/LFM2.5-1.2B-Instruct using speculative decoding in SGLang.
  • On-device inference on Apple silicon via the Metal backend as a draft model for the target LFM2.5-1.2B-Instruct model.
  • MATH-500 and GSM8K style mathematical reasoning tasks when used with the compatible target model, based on reported acceptance benchmarks.
  • HumanEval and MBPP style code generation tasks when used with the compatible target model, based on reported acceptance benchmarks.
  • MT-Bench style multi-turn conversation when used with the compatible target model, based on reported acceptance benchmarks.

Tradeoffs

  • Must be used with the specific target model LiquidAI/LFM2.5-1.2B-Instruct; it is a draft model and not a standalone generation model.
  • Requires a build of SGLang with DSpark support; compatibility with other inference frameworks is not stated.
  • The DSpark family includes separate drafters for different target models; this model is not interchangeable with other targets.
  • License is lfm1.0 (other), and no additional permissions or restrictions beyond the license file are specified in the provided facts.
  • The provided facts do not include quality benchmarks for the draft model itself; generated text is verified by the target model, so performance depends on the target.

Ways to run it

SGLang · Linux

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

python -m sglang.launch_server --model-path LiquidAI/LFM2.5-1.2B-Instruct --speculative-algorithm DSPARK --speculative-draft-model-path LiquidAI/LFM2.5-1.2B-Instruct-DSpark --speculative-draft-attention-backend flashinfer --disable-radix-cache --mem-fraction-static 0.75 --port 30000
Official instructions

Explore its uses

SAFETENSORSTEXT-GENERATION

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 82 out of 100. Release age: 53 days old, Sep 17, 2026. Measures age, not quality.
Popularity 85 out of 100. 1,688,388 downloads in 30 days; 2,960 lifetime likes. Checked Nov 8, 2026. Measures adoption, not quality.
MODEL SIZE27.36B
LOCALRENTED GPU
Released Sep 17, 2026
Text
DeepSeek

DeepSeek V4.1 Flash

An open-weight model for reasoning, coding, agents, and image understanding. DeepSeek V4.1 Flash combines a 552B backbone with sparse activation and up to one million tokens of context. Choose the original checkpoint or a community uncensored variant below.

Freshness 79 out of 100. Repository age: 60 days old, Sep 10, 2026. Measures age, not quality.
Popularity 86 out of 100. 1,842,868 downloads in 30 days; 4,898 lifetime likes. Checked Nov 8, 2026. Measures adoption, not quality.
MODEL SIZE552B backbone · 8B active prefill / 16B active decode
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 →