IPEC-COMMUNITY (Shanghai AI Laboratory affiliated)
EO-1
3.77B Commercial use OK 8/8 verified
Facts
- Released
- 2025-08 1 Source
- Parameters
- 3.77 B 1 Source
- Architecture
- Monolithic autoregressive 0 Source
- License
- MIT (model card) 1 Source
- Commercial
- Commercial use OK 0 Source
- Weights
- huggingface.co ↗ 1 Source
- Code
- github.com ↗ 1 Source
- Paper
- arxiv.org ↗ 1 Source
- Input modalities
- rgb · language · state 0 Source
- Action space
- Joints/end-effector (flow matching) 0 Source
- Cross-embodiment
- AgiBotWorld + Open X-Embodiment + RoboMIND + SO100 community data; Franka Panda / WidowX 250S / AgiBot G-1 / SO100 1 Source
- Pretraining data (hours)
- Unverified 0 Source
- Target forms
- Single arm · Dual arm 0 Source
Architecture · Single decoder Transformer unifying discrete autoregression + continuous flow matching; Qwen2.5-VL-3B backbone; EO-Data1.5M interleaved data
Sources · 3 sources
arch_notes,cross_embodiment,code_url https://github.com/IPEC-PUBLIC/EO-1 Official page · Fetched 2026-09-04 · No snapshot stored yet README：Qwen2.5-VL-3B-Instruct；EO-Data1.5M；Franka Panda, WidowX 250 S, AgiBot G-1, LeRobot SO100
params_b,weights_url,license,release_date https://huggingface.co/api/models/IPEC-COMMUNITY/EO-1-3B Model hub page · Fetched 2026-09-04 · No snapshot stored yet safetensors 3,771,607,072；license: MIT；createdAt 2025-08-28
paper_url https://arxiv.org/abs/2508.21112 Paper / official docs · Fetched 2026-09-04 · No snapshot stored yet EO-1: An Open Unified Embodied Foundation Model for General Robot Control
Activity signals
Public counters fetched daily from GitHub, Hugging Face and ModelScope. Numbers only; no judgement. · fetched 2026-10-31
GitHub stars
0
±0 / 30 d
Commits · 90 d
0
last push 403 d ago
HF downloads · 30 d
65
all-time 2,762 · 15 likes
Weights last modified
2026-01-14
291 d ago
ModelScopeno same-name mirror found (searched by repo name)
Hugging FaceIPEC-COMMUNITY/EO-1-3B
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 | 98.2 | paper · Table 3, row 'EO-1 (Ours)', LIBERO-Overall SR |
| LIBERO | goal | 99.2 | paper · Table 3, row 'EO-1 (Ours)', LIBERO-Goal SR |
| LIBERO | long | 94.8 | paper · Table 3, row 'EO-1 (Ours)', LIBERO-Long SR |
| LIBERO | object | 99.8 | paper · Table 3, row 'EO-1 (Ours)', LIBERO-Object SR |
| LIBERO | spatial | 99.7 | paper · Table 3, row 'EO-1 (Ours)', LIBERO-Spatial SR |
| Other benchmarksEO-1 (3B) | EO-Bench @ Spatial | 36.4 score | paper · Table 2, row 'EO-1 (3B)', column 'EO-Bench @ Spatial' |
| Other benchmarksEO-1 (3B) | EO-Bench @ Temporal | 38.9 score | paper · Table 2, row 'EO-1 (3B)', column 'EO-Bench @ Temporal' |
| Other benchmarksEO-1 (3B) | ERQA | 45.5 score | paper · Table 2, row 'EO-1 (3B)', column 'ERQA' |
| Other benchmarksEO-1 (3B) | Overall | 44.8 score | paper · Table 2, row 'EO-1 (3B)', column 'Overall' |
| Other benchmarksEO-1 (3B) | RoboVQA | 58.5 score | paper · Table 2, row 'EO-1 (3B)', column 'RoboVQA' |
| Real-robot evaluation (self-reported) | AgiBot G-1 long-horizon avg completion | 81 | paper · Section 4.6 text describing Figure 7: 'EO-1 achieves an impressive average completion score of 81.0%' |
| Real-robot evaluation (self-reported) | diverse manipulation, all platforms avg | 86 | paper · Section 4.5 text describing Figure 6: 'achieving a performance of 86.0%, compared with 43.0% for Fast, 71.0% for GR00T-N1.5, and 68.0% for pi0' |
| SimplerEnv · Google Robot | variant_aggregation | 63 | paper · Table 4, Google Robot Benchmark (Aggregation), row 'EO-1 (Ours)', Average |
| SimplerEnv · Google Robot | visual_matching | 76.5 | paper · Table 4, Google Robot Benchmark (Matching), row 'EO-1 (Ours)', Average |
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
arch_notes,cross_embodiment,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 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EO-1 |
Legend Officially supportedCommunity reproducedFine-tunable in theoryUnsupportedUnknown Vendor-declared target form 5 / 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.