Specialist modelOpen weights
LFM2.5-VL-3B-DSpark
Liquid AI
- Released
- September 24, 2026
- Data date
- September 26, 2026
- Model class
LFM2.5-VL-3B-DSpark is a speculative-decoding draft model. It proposes candidates for the target model LFM2.5-VL-3B, which verifies and corrects them when needed. The draft and target model form one inference system. DSpark accelerates local decoding for the target model, with open weights aimed at runtimes such as SGLang and MLX-VLM.
Liquid AI measures up to 2.66x decoding on an H100 and up to 3.13x on an Apple M5 Max under specified runtimes and settings. Image encoding and prefill do not become faster, so end-to-end gains can be smaller.
Specifications and access
| Specification | Value and source |
|---|---|
| Model class | Speculative-decoding draft modelSource |
| Model ID | LiquidAI/LFM2.5-VL-3B-DSparkSource |
| Target model | LiquidAI/LFM2.5-VL-3BSource |
| Access | Local download, no standalone hosted API serviceSource |
| License | LFM 1.0Source |
| Speed | Up to 2.66x faster decoding on H100 and 3.13x on Apple M5 Max, both vendor-measuredSource |
| Benchmark note | Quality benchmarks belong to the target model. DSpark measurements concern decoding and end-to-end latency on specified hardware and runtimes.Source |
Popularity
29.7 / 100
Sources and data date
| Type | Evidence and data date |
|---|---|
| Additional source | Liquid AI model card (retrieved September 26, 2026) · Liquid AI model card · Editorial description reviewed October 4, 2026 |
| Additional source | Liquid AI announcement (retrieved September 26, 2026) · LFM2.5-VL-DSpark announcement · Editorial description reviewed October 4, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |