Shanghai AI Laboratory InternRobotics
InternVLA-M1
3.75B Non-commercial 8/8 verified
Facts
- Released
- 2025-09 1 Source
- Parameters
- 3.75 B 1 Source
- Architecture
- VLM + action expert 0 Source
- License
- Code MIT; weights model card tagged CC BY-NC-SA 4.0 2 Source
- Commercial
- Non-commercial 0 Source
- Weights
- huggingface.co ↗ 1 Source
- Code
- github.com ↗ 1 Source
- Paper
- arxiv.org ↗ 1 Source
- Input modalities
- rgb · language 0 Source
- Action space
- End-effector pose delta + gripper 0 Source
- Cross-embodiment
- RT-1 / Bridge pretraining, LIBERO and SimplerEnv fine-tuning 0 Source
- Pretraining data (hours)
- Unverified 0 Source
- Target forms
- Single arm 0 Source
Architecture · Qwen2.5-VL-3B backbone, spatial guidance + multimodal multi-task joint training
Sources · 3 sources
license,arch_notes,code_url https://github.com/InternRobotics/InternVLA-M1 Official page · Fetched 2026-09-04 · No snapshot stored yet README: MIT license (code); Qwen2.5-VL backbone; 0918 Release model weights
params_b,weights_url,license,release_date https://huggingface.co/api/models/InternRobotics/InternVLA-M1 Model hub page · Fetched 2026-09-04 · No snapshot stored yet safetensors 3,754,622,976；license: cc-by-nc-sa-4.0；createdAt 2025-09-16；base Qwen2.5-VL-3B-Instruct
paper_url https://arxiv.org/abs/2510.13778 Paper / official docs · Fetched 2026-09-04 · No snapshot stored yet InternVLA-M1 technical report
Activity signals
Public counters fetched daily from GitHub, Hugging Face and ModelScope. Numbers only; no judgement. · fetched 2026-10-31
GitHub stars
436
+4 / 30 d
Commits · 90 d
2
last push 23 d ago
HF downloads · 30 d
1,091
all-time 7,478 · 28 likes
Weights last modified
2025-10-15
382 d ago
GitHubInternRobotics/InternVLA-M1 · 26 forks · 9 open issues · no GitHub release
Hugging FaceInternRobotics/InternVLA-M1
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 | 95.9 | paper · Table 4, row 'InternVLA-M1', column Avg |
| LIBERO | goal | 93.8 | paper · Table 4, row 'InternVLA-M1', column Goal |
| LIBERO | long | 92.6 | paper · Table 4, row 'InternVLA-M1', column Long |
| LIBERO | object | 99 | paper · Table 4, row 'InternVLA-M1', column Object |
| LIBERO | spatial | 98 | paper · Table 4, row 'InternVLA-M1', column Spatial |
| Other benchmarksInternVLA-M1 (3B) | Desktop Sorting | 91 | paper · Table 5 / Figure 12, row 'Ours-3B', column 'Desktop Sorting' |
| Other benchmarksInternVLA-M1 (3B) | Goods Purchase | 92 | paper · Table 5 / Figure 12, row 'Ours-3B', column 'Goods Purchase' |
| Other benchmarksInternVLA-M1 (3B) | Make Sandwiches | 91 | paper · Table 5 / Figure 12, row 'Ours-3B', column 'Make Sandwiches' |
| Other benchmarksInternVLA-M1 (3B) | Math calculation | 93 | paper · Table 5 / Figure 12, row 'Ours-3B', column 'Math calculation' |
| Other benchmarksInternVLA-M1 (3B) | Sort into Drawers | 90 | paper · Table 5 / Figure 12, row 'Ours-3B', column 'Sort into Drawers' |
| SimplerEnv · Google Robot | variant_aggregation | 76 | paper · Table 1, row 'InternVLA-M1', Variant Aggregation Avg |
| SimplerEnv · Google Robot | visual_matching | 80.7 | paper · Table 1, row 'InternVLA-M1', Visual Matching Avg |
| SimplerEnv · WidowX (Bridge) | overall | 71.7 | paper · Table 2, row 'InternVLA-M1', Avg |
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
license,arch_notes,code_url · params_b,weights_url,license,release_date · paper_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 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| InternVLA-M1 |
Legend Officially supportedCommunity reproducedFine-tunable in theoryUnsupportedUnknown Vendor-declared target form 4 / 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.