Public KG record
LFM2.5-VL-3B
Recorded overview
A 3B edge-optimized vision-language model pairing SigLIP2 400M with a 2.6B text
An evidence-backed entity from the project knowledge graph; detailed tool metadata is still being added.
Recorded details
- Benchmarks: BFCL v4
- Benchmarks: RefCOCO-avg
- Benchmarks: ScreenSpot-v2
- Benchmarks: ToolSandbox
- Capabilities: 仅保留自身原本会生成的token
- Capabilities: 在贪心解码下生成与基础模型相同的输出
- Capabilities: 函数调用
- Capabilities: 物体定位
Evidence coverage
The KG currently records 9 known information dimensions and has 9 undocumented dimensions, linked to 5 public source articles. These counts describe record coverage, not model performance or verified fit.
Related public records
- Liquid AI — develops
- d1-3B — based on
- Gemma-4-E2B — competes with
- Gemma-4-E4B — competes with
- InternVL-3.5-4B — competes with