Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #8436

AppWorld

Related event timeline, sources and context from the news index.

EVENT TIMELINE

4

RESEARCH · 1 SOURCE · Hugging Face

IBM's CUGA harness improves AppWorld consistency with GLM-5.3-flash versus Hermes and Deep Agents

Researchers ran the same open-weight model (GLM-5.3-flash) across three agent harnesses—IBM's CUGA, Hermes (Nous Research), and LangChain's Deep Agents—on 90 AppWorld tasks; task completion varied from about 30% to 81% depending on the harness, and scenario-level consistency (SGC) favored CUGA (20 scenarios) over Hermes (8) and Deep Agents (2). CUGA is open-source, built for enterprise deployment (including sovereign and air-gapped environments), and adds a runtime policy layer (described in the authors' ACM CAIS 2026 paper) that enforces approvals and tool guards rather than relying on prompts; the authors note limits: one benchmark, non-identical setups, and no policy-enabled runs in the 90-task comparison.

7.0

RESEARCH · 1 SOURCE · arXiv cs.AI

Agentic World Modeling (AWM) enables parallel agent test-time scaling without replicating environments (arXiv:2610.40927v1)

This new preprint introduces Agentic World Modeling (AWM), which casts next-observation prediction as a multi-turn evidence-acquisition task: for each hypothetical action an agentic simulator issues adaptive read-only queries to the real environment, reasons over returned evidence, and uses executable computation to produce observations without mutating shared state. Integrated with trajectory-level test-time scaling, AWM enables parallel exploration over a single shared, non-resettable environment and reportedly outperforms a single-shot simulation baseline by 40.07% on average and a trained world-model baseline by 12.36% on average across τ-bench, AppWorld, and OfficeBench.

7.0

RESEARCH · 1 SOURCE · arXiv cs.AI

Paper proposes using small local models as interpreters for cloud-written agent programs

The arXiv preprint (arXiv:2610.14209v1) studies a hybrid agent architecture where a cloud-written program defers bounded, runtime-dependent questions over private user data to a small local model via function calls. Evaluated on AppWorld across six backbone/test splits, this approach improved code-only agent success rates in all cases—by up to 14.9 percentage points—closing much of the gap to a fully cloud-hosted ReAct loop while keeping data on-device.

6.0

RESEARCH · 1 SOURCE · arXiv cs.AI

Privileged Self-Practice (PSP) for multi-turn LLM agents (arXiv:2609.29051v1)

This arXiv preprint identifies a failure mode of on-policy self-distillation (OPSD) in multi-turn agents—training yields overconfident behavior without the underlying information—and proposes Privileged Self-Practice (PSP). PSP keeps privileged information (PI) in the prompt/sampler rather than the loss: when a student fails rollouts, an analyzer model injects a short per-task instruction into the prompt, the task is re-sampled, and the result is trained with the same GRPO objective; across AppWorld and SWE-bench Verified and three student models PSP consistently outperforms OPSD and plain GRPO, improving task-goal completion by up to 65% on AppWorld and resolved rate by up to 61% on SWE-bench Verified.

7.0