LeRobot releases RynnValue-4B reward model and LeRobot 0.7 with interactive rollouts
Overview
On Nov 5, LeRobot released the Apache-2.0 licensed RynnValue-4B model for predicting task completion time. On Nov 9, LeRobot 0.7 launched interactive rollouts enabling mid-run steering via FineART-VLA.
Generated from attributed reports · Updated 15 hours ago
Event evidence and corrections
1 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Dataset · license: apache-2.0
“model in Python from lerobot.configs.rewards import RewardModelConfig from lerobot.rewards import make_reward_model config = RewardModelConfig.from_pretrained( "lerobot/RynnValue-4B" ) reward_model = make_reward_model(config) # `batch` is the output of this model's preprocessor. remaining_time_s = reward_model.predict_remaining_time(batch).remaining_time_s Model Details License: apache-2.0”
Exact source · revision 2huggingface
Developments
- 2026-11-09 10:27 UTC · 1 reportsLeRobot 0.7 Interactive Rollouts ReleaseLeRobot — X：LeRobot 0.7: Interactive Rollouts for Mid-Run Steering
- 2026-11-05 08:44 UTC · 1 reportsRynnValue-4B Robotic Value Model ReleaseLeRobot — Model repository updates：LeRobot releases RynnValue-4B reward model for predicting robotic task completion time
Report timeline
Follow attributed reports and material updates.
- LeRobot — XSignalLeRobot 0.7: Interactive Rollouts for Mid-Run Steering
LeRobot 0.7 releases interactive rollouts allowing users to steer robot policies mid-run without reset. Features include /subtask switching, /vqa visual queries, /autosteer goal planning, and episode recording. Any VLM (transformers, vLLM, or inference providers) can serve as a planner. FineART-VLA, trained on 1,718 hours of data with 533,913 subtask labels from scale_AI and Hugging Face, verbalizes next subtasks before execution.
- LeRobot — Model repository updatesSignalLeRobot releases RynnValue-4B reward model for predicting robotic task completion time
LeRobot has released RynnValue-4B, a language-conditioned robotic value model built on Qwen3-VL that predicts the remaining time in seconds until a task is completed given a trajectory video and task description. The model is inference-only within the LeRobot ecosystem and cannot be trained using lerobot-train. It is available under an Apache-2.0 license.
Event coverage history
Current attention 7·Peak within the comparable range 10（2026-11-09 12:00 UTC）·Change within the comparable range over 24 hours –
The trend compares participants observed continuously; its scope may be narrower than current attention totals. Hover or click to explore hourly observations, or use the left and right arrow keys.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).