Improving a pretrained vision-language-action (VLA) model requires demonstrations that address its weaknesses and provide effective supervision for learning. Coding agents can autonomously generate demonstrations through environment interaction, yet successful execution alone does not ensure effective supervision for VLA learning. We introduce Agentic-Teacher, an autonomous data engine for failure-driven VLA post-training. Given a VLA checkpoint, a task specification, and configured environment tools, the engine analyzes failed rollouts, selects intervention states, and iteratively develops closed-loop code policies through execution feedback. Validated policies become reusable demonstration generators that collect successful continuations from states where the VLA struggles. To make these demonstrations suitable for learning, Agentic-Teacher introduces temporally correlated action noise to broaden state-action coverage, removes idle frames while preserving gripper transitions, and filters trajectories using the frozen base model's flow-matching loss. The resulting demonstrations are combined with successful VLA rollouts for post-training, yielding an updated policy that operates without coding-agent assistance. Agentic-Teacher thus connects the weaknesses of a target VLA to both the generation and curation of its training data. In simulation, it improves π0 success rates from 79.4% to 87.8% on LIBERO-10, from 80.4% to 90.6% on MetaWorld-10, and from 0% to 30.0% across three RoboCasa tasks.
Figures & tables
Figure 1: Agentic-Teacher pipeline. Given a VLA checkpoint, simulation environment, and task interface, a coding agent uses the harness to evaluate the VLA under varied initial conditions and diagnose failures by comparing successful and failed rollouts. It invokes perception and control tools to synthesize and validate closed-loop code policies that generate corrective demonstrations from VLA failure states. Demonstrations from reliable repair experts are processed through action noise injection, idle-frame removal, and filtering based on the frozen base model’s flow-matching loss, and then combined with successful VLA trajectories to construct the VLA fine-tuning dataset.
Figure 2: A VLA grasp failure and the corresponding code-policy repair. (a) The VLA approaches off target, misses the grasp, and fails to lift the bowl. (b) The repair program realigns the gripper, grasps the bowl, and lifts it. Columns show selected rollout steps, with an additional camera view in each inset.
Figure 3: Motion and action statistics of initial code-policy trajectories, processed training episodes, and successful VLA rollouts. (a–c) Selected end-effector paths. (d) Distributions of translation and rotation action magnitudes in OSC units, with dots showing episode medians and numbers showing episode-weighted means. (e) Fraction of near-zero transitions. (f) Flow matching loss; dots denote episode scores and diamonds denote group means.
Figure 4: VLA success rates across benchmarks.
Task
Failure mode
Before repair (%) ↓
After repair (%) ↓
Put the cream cheese box and butter in the basket.
Butter approach obstruction
8.0
2.0
Basket-rim collision
4.0
0.0
Cheese slippage
2.0
0.0
Put both moka pots on the stove.
First-pot setdown rotation
8.0
0.0
Second-pot approach obstruction
10.0
8.0
Handle misalignment
22.0
0.0
Table 1: Per-task failure-mode incidence before and after repair on three selected LIBERO-10 tasks. All rates use the corresponding task’s 50 evaluation trials as the denominator.
Figure 5: Cumulative trajectory-processing ablation on LIBERO at training step 4,000. Each successive variant adds the indicated processing step to the preceding variant. Panels show (a) task-average success rates, (b) trajectory counts, and (c) retained-frame counts.
Figure 6: (a) Agentic-Teacher versus the reported PPO baseline, averaged across RoboCasa tasks. (b) Agentic-Teacher repair efficiency on individual RoboCasa and LIBERO tasks.
Figure 7: Repair coverage with and without successful references: (a) drawer opening; (b) pick-and-place.
Appendix figures & tables4 assets
Supplementary material from the paper’s appendix.
Appendix
Parameter
MetaWorld MT10
LIBERO-10
RoboCasa
Initialization
RLinf π0 MetaWorld SFT
Pre-repair π0 LIBERO checkpoint
π0 human300 multitask, step 75,000
Peak learning rate
1×10−5
5×10−5
2.5×10−5
Global batch size
32 (8 per GPU)
32 (8 per GPU)
64
Action horizon
5
20
20 / 50
Training steps
20,000
10,000
5,000 / 10,000
Warmup steps
1,000
200
1,000
Appendix
Table 2: Supervised post-training configurations. RoboCasa entries list the action horizons and training budgets used across its experiment configurations.
ID
Task
Base π0 (%)
Agentic-Teacher (%)
0
soup-tomato
88
94
1
cheese-butter
86
98
2
stove-moka
86
96
3
bowl-drawer
96
98
4
mugs-plates
76
90
5
book-caddy
96
88
Appendix
Table 3: LIBERO-10 task success rates before and after Agentic-Teacher post-training. Each policy is evaluated with 50 rollouts per task. The average gives equal weight to each task.
Task
Base π0 (%)
Agentic-Teacher (%)
reach-v3
40
84
push-v3
58
84
pick-place-v3
48
70
door-open-v3
100
100
drawer-open-v3
100
98
drawer-close-v3
100
100
Appendix
Table 4: MetaWorld-10 task success rates before and after Agentic-Teacher post-training. Each policy is evaluated with 50 rollouts per task. The average gives equal weight to each task.
Task
Base π0 (%)
Agentic-Teacher (%)
CloseToasterOvenDoor
0
48
VeggieDipPrep
0
30
ClearCuttingBoard
0
12
Average
0.0
30.0
Appendix
Table 5: RoboCasa task success rates before and after Agentic-Teacher post-training. Each policy is evaluated with 50 rollouts per task. The average gives equal weight to each task.
School of computing and data science, Nanyang Technological University · School of Software Engineering, Sun Yat-sen University, China and GuangDong Engineering Technology Research Center of Blockchain, China · Northwest A&F University