Midea AI Research et al. (DexVLA team)
DexVLA
3B Non-commercial 8/8 verified
Facts
- Released
- 2025-02 1 Source
- Parameters
- 3 B 0 Source
- Architecture
- VLM + action expert 0 Source
- License
- Code MIT; ScaleDP pretrained weights model card tagged CC BY-NC 4.0 2 Source
- Commercial
- Non-commercial 0 Source
- Weights
- huggingface.co ↗ 1 Source
- Code
- github.com ↗ 1 Source
- Paper
- arxiv.org ↗ 0 Source
- Input modalities
- rgb · language · state 0 Source
- Action space
- Dual-arm / dexterous-hand joints (diffusion) 0 Source
- Cross-embodiment
- ALOHA scripts, AgileX evaluation; README lists no training hours 0 Source
- Pretraining data (hours)
- Unverified 0 Source
- Target forms
- Dual arm · form.dexterous_hand 0 Source
Architecture · Qwen2-VL-2B backbone + plug-in ScaleDP diffusion expert (1B or 410M)
Sources · 2 sources
release_date,arch_notes,license,code_url https://github.com/juruobenruo/DexVLA Official page · Fetched 2026-09-05 · No snapshot stored yet README: released 2025-02-17; Qwen2-VL-2B + 1B ScaleDP (also 410M); MIT LICENSE file
weights_url,license https://huggingface.co/api/models/lesjie/scale_dp_h Model hub page · Fetched 2026-09-05 · No snapshot stored yet Model card license: cc-by-nc-4.0; createdAt 2025-02-24
Activity signals
Public counters fetched daily from GitHub, Hugging Face and ModelScope. Numbers only; no judgement. · fetched 2026-10-31
GitHub stars
73
±0 / 30 d
Commits · 90 d
0
last push 565 d ago
HF downloads · 30 d
0
all-time 0 · 0 likes
Weights last modified
2025-02-24
615 d ago
ModelScopeno same-name mirror found (searched by repo name)
Hugging Facelesjie/scale_dp_h
stars · last 30 d
Public benchmark scores
As printed in the source; protocols differ between papers. Full table with notes on the benchmarks page.
| Benchmark | Subset | Score | Source |
|---|---|---|---|
| LIBERO | avg | 97.3 | paper · Table 3, row 'DexVLA', column 'Average' |
| LIBERO | goal | 95.6 | paper · Table 3, row 'DexVLA', column 'Goal' |
| LIBERO | object | 99.1 | paper · Table 3, row 'DexVLA', column 'Object' |
| LIBERO | spatial | 97.2 | paper · Table 3, row 'DexVLA', column 'Spatial' |
| Real-robot evaluation (self-reported) | laundry_folding_score | 0.4 score 0-1 | paper · Table 1, row Stage1+Stage2+Stage3 (all checked), column 'Laundry Folding' |
| Real-robot evaluation (self-reported) | shirt_folding_score | 0.9 score 0-1 | paper · Table 1 'Ablation study of three stage training', row Stage1+Stage2+Stage3 (all checked), column 'Shirt Folding'; also text Sec. 4: 'DexVLA achieves a 0.92 point on shirt folding without any task-specific adaptation' |
Deployment recipe
No recipe yet. We publish only steps that actually ran on a rented GPU with a measured latency; this model has not been through that yet.
Record history
2026-09-05
Sources fetched
release_date,arch_notes,license,code_url · weights_url,license
Latency & cost
cost = p50 ms ÷ 1000 × rental ÷ 3600 × 1000
| Hardware | p50 ms | p95 ms | Peak VRAM | Cost per 1k inferences ¥ | samples |
|---|---|---|---|---|---|
NVIDIA RTX 4090 ¥1.68/h · GongJi Compute (2026) | — | — | — | — no latency data | No data |
NVIDIA RTX 5090 ¥2.4/h · Runyun 2.4 / GongJi 2.5 / Paratera Cloud 2.98 (2026) | — | — | — | — no latency data | No data |
NVIDIA A100 80GB no rental price | — | — | — | — no latency data | No data |
NVIDIA Jetson AGX Thor no rental price | — | — | — | — no latency data | No data |
NVIDIA H100 80GB no rental price | — | — | — | — no latency data | No data |
NVIDIA A800 80GB no rental price | — | — | — | — no latency data | No data |
No latency measurements for this model yet · Data comes from literature, maintainer runs, or crowd CLI uploads. None exist so far.
Compatibility
Colored only where we hold evidence. An amber dot means the vendor declares this form as a target; that is a claim, not a verified compatibility.
| Unitree G1 | Unitree H1 | AgiBot Genie G1 | AgileX Piper arm | SO-101 | Franka FR3 / Panda | Fourier GR-2 | WidowX 250 S | Google Robot（Everyday Robots） | ALOHA 2 / Cobot Magic (dual-arm) | Galaxea R1 Pro | Franka Emika Panda | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DexVLA |
Legend Officially supportedCommunity reproducedFine-tunable in theoryUnsupportedUnknown Vendor-declared target form 3 / 12 cells with a definite status
Reproduction records
Did this model run on your robot? Report it with a link (log, video or repo). Records are reviewed by the maintainers; a verified “works” or “works after fine-tuning” upgrades the matrix cell to “community verified” with your link as evidence. Anonymous is fine; nothing but the fields below is stored.
No verified record yet for this model.
Clear direction, grounded judgement.