Robbyant (Ant Group)
LingBot-VLA 4B
4.2B Commercial terms unverified 7/8 verified
1 of 8 required fields are unverified. They are shown as badges, not filled with defaults. Sources for verified fields are listed below.
Facts
- Released
- 2026-01 1 Source
- Parameters
- 4.2 B 1 Source
- Architecture
- VLM + action expert 0 Source
- License
- Apache-2.0 (code repository); weights model card has no licence tag 1 Source
- Commercial
- Commercial terms unverified 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 joints 0 Source
- Cross-embodiment
- 9 dual-arm configurations, 20k hours of real-robot data; measured on AgiBot G1 / AgileX / Galaxea R1Pro 1 Source
- Pretraining data (hours)
- 20000 h 1 Source
- Target forms
- Dual arm · Dual arm, wheeled 0 Source
Architecture · Qwen2.5-VL-3B-Instruct backbone + MoGe-2 depth component; depth-free and depth-distilled versions; 20k hours of real-robot dual-arm data
Sources · 2 sources
release_date,license,arch_notes,cross_embodiment,pretrain_hours,code_url https://github.com/robbyant/lingbot-vla Official page · Fetched 2026-09-04 · No snapshot stored yet README: technical report, weights and code released 2026-01-27; Apache-2.0 License; Qwen2.5-VL-3B-Instruct; 20,000 hours of real-world data from 9 dual-arm configurations; Agibot G1 / AgileX / Galaxea R1Pro
params_b,weights_url https://huggingface.co/api/models/robbyant/lingbot-vla-4b Model hub page · Fetched 2026-09-04 · No snapshot stored yet safetensors 4,197,425,739; model card has no license field; createdAt 2026-01-26
Activity signals
Public counters fetched daily from GitHub, Hugging Face and ModelScope. Numbers only; no judgement. · fetched 2026-10-31
GitHub stars
1,882
+43 / 30 d
Commits · 90 d
0
last push 143 d ago
HF downloads · 30 d
876
all-time 7,282 · 33 likes
Weights last modified
2026-04-30
185 d ago
GitHubrobbyant/lingbot-vla · 192 forks · 41 open issues · no GitHub release
Hugging Facerobbyant/lingbot-vla-4b
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 |
|---|---|---|---|
| Real-robot evaluation (self-reported)LingBot-VLA-4B-Depth (w/ depth distillation) | gm100_3_platforms_avg_success_rate | 17.3 | official · README GM-100 table (3 platforms: Agibot G1, AgileX, Galaxea R1Pro), column 'Ours w/ depth', Average SR / PS = 17.30% / 35.41% |
| Real-robot evaluation (self-reported)LingBot-VLA-4B-Depth (w/ depth distillation) | gm100_4_platforms_avg_progress_score | 35.2 % progress | paper · GM-100 table, column 'Ours w/ depth', Average PS |
| Real-robot evaluation (self-reported)LingBot-VLA-4B (w/o depth) | gm100_4_platforms_avg_success_rate | 16.2 | paper · GM-100 table, column 'Ours w/o depth', Average SR (PS 34.32%) |
| Real-robot evaluation (self-reported)LingBot-VLA-4B-Depth (w/ depth distillation) | gm100_4_platforms_avg_success_rate | 16.9 | paper · GM-100 multi-platform table, column 'Ours w/ depth', Average SR (Progress Score 35.16%) |
| RoboTwin 2.0LingBot-VLA-4B (w/o depth) | clean | 86.5 | paper · RoboTwin 2.0 table, column 'Ours w/o depth', Clean Scenes (Randomized 85.34) |
| RoboTwin 2.0LingBot-VLA-4B-Depth (w/ depth distillation) | clean | 88.6 | paper · RoboTwin 2.0 table, column 'Ours w/ depth', Clean Scenes |
| RoboTwin 2.0LingBot-VLA-4B-Depth (w/ depth distillation) | randomized | 86.7 | paper · RoboTwin 2.0 table, column 'Ours w/ depth', Randomized Scenes |
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
2026-09-04
Sources fetched
release_date,license,arch_notes,cross_embodiment,pretrain_hours,code_url · params_b,weights_url
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 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LingBot-VLA 4B |
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.