Robot makerBeijing, ChinaFounded 2023
Builds wheeled and bipedal humanoid robots and, with Peking University, trains task-specific models such as GraspVLA, TrackVLA and NavFoM.
Galbot is a Beijing-based embodied AI and humanoid robotics company founded in May 2023 by Wang He, an assistant professor at Peking University, with co-founder Yao Tengzhou, formerly of ABB's robotics R&D center. It develops general-purpose robots and embodied multimodal models. In March 2026 it raised RMB 2.5 billion and was reported to be preparing a Hong Kong IPO.
Sources: news.cn (23 Jun 2025) · 21jingji.com (19 Sep 2024) · tmtpost.com (23 Jun 2025) · caixinglobal.com (3 Mar 2026)
Private company · $662M disclosed in USD across 2 of 5 listed rounds
| Date | Round | Amount | Lead | Source |
|---|---|---|---|---|
| Jun 2024 | Angel | RMB 700M | n/a | 21jingji.com, 19 Sep 2024 |
| Nov 2024 | Strategic | RMB 500M | n/a | news.cn, 23 Jun 2025 |
| 23 Jun 2025 | Strategic | RMB 1.1B | CATL | news.cn, 23 Jun 2025 |
| Dec 2025 | UndisclosedValuation: $3B post-money | Over $300M | China Mobile-backed fund | caixinglobal.com, 3 Mar 2026 |
| 2 Mar 2026 | Undisclosed | RMB 2.5B (about $362M) | n/a | caixinglobal.com, 3 Mar 2026 |
#53 of 104 model companies · score 34.2 / 100 · updated 19 Oct 2026
Each bar shows an input from 0 to 100% before weighting. 1 of 3 listed models have open weights. Most-starred official repository: ShaoanWang/TrackVLA (437 stars, 19 Oct 2026).
The score (0–100) measures public, checkable evidence about a company's robot models. It does not measure private capability. A company that publishes no benchmark results and no weights can still have strong models. Every input is computed from the data on each company page, and every number there links to its source.
| Input | Weight | How it is measured |
|---|---|---|
| Published benchmarks | 35% | For each scored benchmark, the company's best result divided by the best result any company has published, averaged over the benchmarks it reports. No published results scores 0. Results come only from the model's own paper, model card or the official benchmark leaderboard. Scored: LIBERO (4-suite average success, %); SimplerEnv Google Robot, visual matching (%); SimplerEnv Google Robot, variant aggregation (%); SimplerEnv WidowX / Bridge (%); CALVIN ABC→D (average length, 0–5); RoboArena (pairwise rating). |
| Open weights | 20% | Share of the company's listed models whose weights the public can download. |
| Adoption | 20% | GitHub stars on the company's most-starred official model repository, on a log scale relative to the most-starred repository in the list. Hugging Face downloads are shown on company pages but not scored, because gated repositories report zero. |
| Released models | 15% | Number of publicly announced, named models, capped at 5. |
| Disclosed funding | 10% | The sum of the funding rounds listed on the company's profile that were disclosed in USD, on a log scale from $10M to the largest sum in the list. Rounds reported only in other currencies are listed but not converted or counted. This input applies only to private startups and robot makers. For big-tech divisions, research labs and listed companies it is n/a, and its weight is shared across the other inputs in proportion. |
Sources. arXiv papers and official model cards for benchmarks and parameters; official leaderboards (RoboArena); the GitHub and Hugging Face APIs for stars and downloads; company announcements and established press for funding. Missing values are shown as n/a, never estimated.
Updated 19 Oct 2026. Benchmark protocols differ between papers, for example in fine-tuning and number of trials, so small gaps between close scores are not meaningful.
Open weightsVLA
Grasping VLA pretrained on the billion-frame synthetic SynGrasp-1B dataset, combining autoregressive perception with flow-matching action generation; weights released 2025-07-25.
No benchmark results on a shared public benchmark have been published for this model.
Weights: n/aPolicy
Cross-embodiment navigation foundation model trained on about eight million navigation samples covering vision-language navigation, object search, tracking and driving.
No benchmark results on a shared public benchmark have been published for this model.
Closed weightsVLA
Vision-language-action model for embodied visual tracking that shares one LLM backbone between target recognition and an anchor-based diffusion trajectory planner; checkpoint not released as of the repo snapshot.
No benchmark results on a shared public benchmark have been published for this model.
