arXiv:2609. 10954v1 Announce Type: new Abstract: Continual world models must decide whether new data justify changing the model.
arXiv:2608. 08239v1 Announce Type: new Abstract: LLM routers promise efficiency by matching each request to the cheapest adequate model, and are increasingly applied per step inside multi-step agents.
arXiv:2609.25052v1 Announce Type: new Abstract: "An agent that writes its conclusions into a store it later retrieves from closes a loop usually reported as one-way contamination. Taking the loop to...
The paper investigates whether self-consensus—stopping a reasoning model when its partial trajectory’s answers agree—can safely reduce inference cost. A large sweep of 3,520 consensus rules on two models and three benchmarks failed to meet predefined safety criteria, while a boundary‑confidence control (DEER) succeeded. The study shows that agreement signals that an answer persists under a fixed probing procedure, not that reasoning has finished, leading to premature stops and missed corrections even when token savings are significant.
The paper investigates how inherited state affects sub-agent performance in multi-agent frameworks, comparing three inheritance policies—Reset, Selective, and Full—across a ladder of Qwen3 models. It finds that reliance on stale state decreases with model capability, but a mid-capability model (Qwen3‑1.7B) exhibits a statistically significant local minimum of net harm, defining a "danger band." Selective handoff consistently improves accuracy over Full, especially within the danger band, while a fixed-threshold router fails on other datasets.
The paper investigates where exactly‑once semantics should be enforced for tool‑using agents—within the model, the agent harness, or the tool contract—by evaluating 25,930 episodes across nine models, three harnesses, two contract variants, and fifteen recovery conditions. Using the LIMBO sandbox, the study shows that when an immediate read‑back is available, frontier models rarely duplicate lost acknowledgements, whereas weaker models do; when read‑back is unavailable, the contract’s idempotency keys explain most duplicate behavior. The authors prove that verification‑only policies cannot guarantee exactly‑once under late commits without bounded in‑flight time, and that waiting only helps when delays are short and predictable. whyItMatters":"The findings clarify that enforcing exactly‑once semantics largely depends on the tool contract and fault type, guiding designers on where to focus reliability mechanisms for LLM agents."
arXiv:2608. 03222v1 Announce Type: cross Abstract: Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates.
arXiv:2607. 25152v1 Announce Type: new Abstract: Long-running autonomous agents plan, act, and judge their own completion without human intervention.
ERRAND is a new method for budgeted maintenance of agent memory that treats revalidation of stored knowledge as a priced errand competing for scarce actions. It uses an errand index that is single‑peaked, allowing certainty in either direction to cost nothing, and repairs by writing new versions rather than deleting old ones. In experiments across two drifting tool‑use worlds, ERRAND outperforms non‑oracle policies, achieving up to 10.0 percentage points improvement over eager revalidation while using only 11.0% of steps, and it self‑terminates when no budget is imposed.
Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates. Failed runs tend to be longer and exhibit redundant exploration or looping, suggesting that some failures may be detectable before completion.
The paper investigates the safety of Code World Models, where a language model generates executable world models that a planner uses. It shows that accepting a model based on sampled transitions only guarantees sample consistency, not full safety, because the probability of missing critical events decays as (1‑r)^N. Experiments on hybrid instruments reveal that omitted mode‑boundaries can severely limit planner performance, and that even sophisticated LLMs (GPT‑5.x) struggle to repair such omissions in higher‑dimensional settings.
The paper introduces a new evaluation protocol called checkpoint handoff to disentangle the contributions of reaching a target state and solving the task in reinforcement learning agents. By cloning states reached by one checkpoint and handing them to another without retraining, the authors separate the REACH metric (how often a policy arrives at a state confirmed to be a fixed number of actions from success) from the SOLVE metric (how often it finishes from that identical state). Across two benchmarks and pipelines, the analysis shows that RL history benefits RL solvers more than SFT solvers, and that independent REACH and SOLVE gaps predict overall performance.
arXiv:2607. 21910v2 Announce Type: replace Abstract: World models let agents plan against predicted physical state, but that state drifts; re-observation is costly and delayed, and repair can fail.