Research map / Vision-language and multimodal / Robot action and navigation
Evo-1: Lightweight Vision-Language-Action Model with Preserved Semantic Alignment
Tao Lin, Yilei Zhong, Yuxin Du, Jingjing Zhang, Jiting Liu, Yinxinyu Chen, Encheng Gu, Ziyan Liu, Hongyi Cai, Yanwen Zou, Lixing Zou, Zhaoye Zhou and 2 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use two-stage training (freeze VLM first) instead of single-stage joint training (ablation) | Evo-1 · Meta-World | success rate | Improved |
| use mid-layer cross-attention integration (Module A) instead of Modules B-D (ablation) | Evo-1 · LIBERO-Long | success rate | Improved |
| keep only first 14 layers of the language branch (ablation) | Evo-1 · design choice | success rate | No clear effect |
| use Evo-1 compact design vs larger SmolVLA on Meta-World | Evo-1 · Meta-World | success rate: 68.2 -> 80.6 | Improved |
| Evo-1 vs pi0 on LIBERO-Spatial | Evo-1 · LIBERO Spatial | success rate: 96.8 -> 92.7 | Hurt |
| Evo-1 vs pi0 on LIBERO average | Evo-1 · LIBERO | success rate: 94.2 -> 94.8 | Improved |
| Evo-1 vs pi0 on real-world tasks | Evo-1 · Real-world xArm6 | success rate: 73 -> 78 | Improved |
| Evo-1 vs SmolVLA under unseen distractor | Evo-1 · Real-world Pick and Place Can | success rate: 65% -> 80% | Improved |
| Evo-1 vs SmolVLA at 30 mm higher target | Evo-1 · Real-world Pick and Place Can | success rate: 60% -> 70% | Improved |
| Evo-1 smaller model vs pi0 on memory | Evo-1 · Inference RTX 4090d | GPU memory (GB): 17.9 -> 2.3 | Improved |
Later papers that compared with this method
2 later accepted papers report Evo-1 as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Spatial Representation Distillation and Knowledge Routing for Vision-Language-Action Models (NeurIPS 2026 · better on 1 of 1 shared result)
- ALAM: Algebraically Consistent Latent Transitions for Vision-Language-Action Models (NeurIPS 2026 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models (NeurIPS 2026 · similarity 0.92)
- Beyond Data Scaling: Representation-Centric Pre-training for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.92)
- VLA-REPLICA: A Low-Cost, Reproducible Benchmark for Real-World Evaluation of Vision-Language-Action Models (NeurIPS 2026 · similarity 0.92)
- PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- OTTER: A Vision-Language-Action Model with Text-Aware Visual Feature Extraction (ICML 2025 · similarity 0.91)
- NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
