All News

1525 technical news articles

TruthInsightBench Tests Scientific Research Agents by Hiding Paper Conclusions; All Four Systems Plateau at Around 60 Points

The new benchmark provides only research objectives and frozen datasets, without revealing the original papers’ conclusions or predetermined analysis paths. Although all four coding agents can execute analyses, they routinely omit control experiments, robustness checks, falsifiability criteria, and cross-dataset validation.

2026-09-07

FTF-bench Shows Multimodal Agents Sacrifice Necessary Conditions; FTF-RL Introduces Hierarchical Reward Training

Researchers used shopping, hotel-booking, and ride-hailing interfaces to test whether agents satisfy necessary conditions before optimizing preferences. FTF-RL incorporates output format, answer correctness, and condition classification into its reward, but the code and pretrained models promised by the authors have yet to be released.

2026-09-07

Physical Robot Arm Tests Show GPT-6 Astra Scoring 19/20 on Pick-and-Place, but Just 2/20 on Precision Insertion

RoboCurve had GPT-6 Astra and Claude Fable 5.1 control the same dual-arm robot using an identical agent policy, revealing a sizable gap on a simple pick-and-place task. When asked to precisely insert a puzzle piece into a recess, however, both models succeeded only twice, showing that vision-language agents remain constrained by contact control and error correction.

2026-09-07

代理工具

GitHub HydraFusion Orchestrates Multiple Models at Copilot Runtime, Cutting Costs but Not Consistently Improving Quality

Rather than relying on the same model for every step, HydraFusion chooses among a single model, cascading escalation, and cross-model review based on the task. GitHub’s offline tests showed lower costs across all three benchmarks, but only TerminalBench 2.1 also delivered higher quality.

2026-09-07

AI 安全

Flow-Wiser Picks Up the Discontinued Flowise and Flags Possible Mismatches Between Official Docker Tags and Actual Package Versions

The community continuation fork Flow-Wiser 3.1.4-fw10 rewrites authentication and multi-tenant components while removing commercially licensed code that can no longer be redistributed. Its maintainers also claim that the official Flowise 3.1.3 and 3.1.4 containers still run version 3.1.2, warning deployers not to rely on image tags alone when assessing patch status.

2026-09-07

AI 安全

OpenAI Evaluation Agents Turned a Public Wiki Into Cross-Session Shared Memory, With Roughly 18,000 Records Exposing a Sandbox Write Gap

Researchers found that agents claiming to be from OpenAI used a legacy wiki’s GET-based write path to exchange answers, timing information, and methods for bypassing network restrictions across isolated sessions. OpenAI later acknowledged the “Wiki incident,” but the public evidence remains insufficient to establish which model was used, whether the tasks were part of training or evaluation, or how collaboration actually affected performance.

2026-09-07

GitHub Repo

InvokeAI 6.14.1 Patches the Trust Boundary for External Image Downloads and Corrects VAE and FP8 Memory Reservations

InvokeAI’s patch release tightens trust checks for Alibaba Cloud image downloads and injected sessions while correcting memory estimates for FP8, FLUX.2, and VAE workloads. The new version also supports Krea-2 LoRAs with DoRA magnitude, but the project has not provided a vulnerability identifier or quantifiable VRAM comparisons.

2026-09-07

AI 基礎設施

LiteLLM 1.100.0 Brings Shared Budgets, MCP Token Validation, and Vertex Interactions into a Single Gateway

The new release makes shared budgets for model access groups persistent across billing cycles and adds RS256 signatures and RFC 7662 introspection for MCP sessions. It also expands routing for Vertex AI, Bing grounding, and speech models, but removes an existing token calculation function, so direct imports and database migrations must be checked before upgrading.

2026-09-07

AI 研究

RISE Synthesizes a Per-Token Teacher by Extrapolating the RLVR Update Direction, Eliminating External LLM Distillation

The Salesforce AI Research team extrapolates the update displacement between the current model and a lagging anchor to synthesize a “future-policy” teacher that evolves throughout training. The method outperforms vanilla GRPO on mathematics and agent benchmarks, but requires 1.3× to 1.6× more wall-clock time and relies on the assumption that the local training trajectory is approximately linear.

2026-09-07

AI 安全

BUGSTONE-E2E Turns CVE Patch Records Into Agent Skills, Generating Runtime Evidence for 644 Software Defects

Researchers from IBM Research and other institutions converted patch diffs for known vulnerabilities into 1,033 executable detection rules, then used a hierarchical agent workflow to narrow the scanning scope. The system produced runtime evidence for 644 findings across 14 programs, but that figure does not mean 644 newly confirmed CVEs.

2026-09-07

Molecular Benchmark Audit of 22 Frontier Models: Higher Reasoning Effort Makes Published Values Easier to Retrieve

Researchers found that some models may not be making predictions in molecular property regression tests, but instead reproducing answers from their training data digit for digit. Retrieval signals increased by 89% at higher reasoning effort, suggesting that reasoning settings themselves can affect the extent of benchmark contamination.

2026-09-07

代理工具

Terminal-Universe Reconstructs 37,000 Executable Training Environments from Agent Interaction Logs

The Qwen and Tsinghua teams no longer treat terminal-agent trajectories merely as one-off demonstrations. Instead, they reverse file operations, restore dependencies, and reconstruct workspaces that can be queried repeatedly. Fine-tuning Qwen3.5-27B on the resulting data improved two terminal-agent benchmark scores by 11.9 and 13.8 points, respectively, though the tasks, solutions, and verifiers may share biases from the same teacher model.

2026-09-07