lab OpenAI News · 2d ago · releases

V7 uses GPT-6 Astra to give AI agents institutional memory

from English

V7 has integrated OpenAI's GPT-6 Astra into its V7 Go platform to provide AI agents with persistent "institutional memory" via a Context Graph, significantly improving retrieval accuracy on complex enterprise queries.

The Context Graph organizes company files into structured entity relationships, allowing agents to query buried context without relearning it for every request. On the hardest graph-query tests, GPT-6 Astra achieved 89% accuracy compared to 78% for GPT-5.6 Sol. V7 also reports that GPT-5.6 Luna reduces document processing costs by 78% and that workflow creation time has dropped from one hour to about 20 minutes.

This approach enables agents to complete 50–100 step workflows in minutes with 99.9% accuracy, helping finance and insurance teams drastically reduce review times and errors.

Importance 1/3 Trust 3/3 OpenAI News OpenAI AI agents Retrieval & RAG

Read original

arxiv arXiv cs.CL · 98d ago · 8 views

ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

ProvenanceGuard introduces a source-aware verifier for MCP-based LLM agents that detects cross-source conflation by routing claims to specific evidence sources and comparing stated attribution with actual source ownership. It achieves block F1 of 0.802 and source accuracy of 0.858 on 260 source-eligible claims, outperforming source-blind baselines, and detects all injected attribution swaps in 50 clinical probes.

arxiv arXiv cs.AI · 98d ago · 7 views

ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

arxiv arXiv cs.CL · 98d ago · 8 views

MODE-RAG: Evaluating and Reducing Hallucinations in M-RAG

MODE-RAG proposes a multi-agent system using Variational Free Energy to dynamically gate interventions and reduce cross-modal hallucinations in retrieval-augmented generation. It integrates Monte Carlo Tree Search and logit perturbations to address causal fabrications and sycophancy, with dedicated agents ensuring factual verification and formatting stability. Evaluated via ModeVent, a subset of MultiVent, the system significantly improves robustness against logical fabrications.

arxiv arXiv cs.AI · 3h ago · 9 views

CliffCompaction reduces coding agent costs by 50% while maintaining performance

Researchers introduce CliffCompaction, an autocompaction technique designed to reduce costs for long-horizon coding agents by up to 50% under a bounded context. The method maintains or improves performance on Terminal-Bench and achieves state-of-the-art results on KernelBench by keeping compacted information faithful through truncation rather than rephrasing.

arxiv arXiv cs.LG · 4h ago · 9 views

CliffCompaction reduces long-horizon coding agent costs by up to 50% while maintaining performance

Researchers introduce CliffCompaction, an autocompaction technique designed for agents handling millions of tokens, which cuts costs by up to 50% under bounded context. The method maintains or improves performance on Terminal-Bench and achieves state-of-the-art results on KernelBench by keeping information faithful through truncation rather than rephrasing.