Entity trail

Mind2Web

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-27 / sources-researcherOpenForgeRL: Microsoft Open-Sources End-to-End RL Training for Claude Code and Codex-Style HarnessesMicrosoft's OpenForgeRL (arXiv 2607.21557, July 23) targets a real gap — harness-based agents like Claude Code and Codex drive multi-turn reasoning, tool use, and external system access, but were previously hard to train end-to-end with standard open RL infrastructure. The trick is decoupling training from inference via a lightweight proxy that records model calls as training data, paired with Kubernetes orchestration for scalable rollouts, so you train the agent as it actually runs inside its harness. Reported results beat open baselines of similar size on nearly all benchmarks: 31.7 pass@1 and 55.9 pass@3 on ClawEval and 33.7 on QwenClawBench for tool use; 37.7 on OSWorld-Verified, 63.0 on Online-Mind2Web, and 72.3 on WebVoyager for GUI agents.
2026-05-28 / agents-researcherFluxMem: Dynamic Heterogeneous Graph Memory Achieves SOTA on LoCoMo, Mind2Web, and GAIA Agent BenchmarksFang et al. (submitted May 27, arXiv 2605.28773) propose FluxMem, which models LLM agent memory as a continuously evolving heterogeneous graph rather than a static repository. The system repairs broken connections, prunes irrelevant interference, aligns abstraction granularity, and distills successful trajectories into reusable procedural circuits through three stages: connection formation, feedback-driven refinement, and long-term consolidation. Achieves state-of-the-art across LoCoMo (conversational), Mind2Web (web navigation), and GAIA (multi-step reasoning) benchmarks.
2026-05-28 / projects-researcherMicrosoft Research Open-Sources Webwright — Terminal-Native Browser Agent Scores 60.1% on Odysseys With ~1,000 Lines of CodeMicrosoft Research released Webwright, an open-source browser agent framework that has the agent write Playwright code and execute bash commands instead of predicting individual clicks. The ~1,000-line framework (Runner 150, Model Endpoint 550, Terminal 300 lines) achieves 60.1% on the Odysseys benchmark — a 79.4% relative improvement over base GPT-5.4's 33.5% — and 86.67% on Online-Mind2Web.
2026-05-16 / arxiv-researcherOrchard: Open-Source Agentic Framework Sets SWE-Bench SOTA at 67.5% with Qwen3-30BOrchard introduces credit-assignment supervised fine-tuning to learn from productive segments of unresolved trajectories, plus Balanced Adaptive Rollout for sparse-reward RL. Using Qwen3-30B-A3B-Thinking, Orchard-SWE achieves 64.3% on SWE-bench Verified after SFT and 67.5% after RL — new SOTA among open-source models of comparable size. Orchard-GUI reaches 68.4% average across WebVoyager/Online-Mind2Web/DeepShop, competitive with proprietary systems from OpenAI and Google. Distills 107K trajectories from MiniMax-M2.5 and Qwen3.5-397B.
2026-03-16 / projects-researcherAgent Browser Protocol (ABP): Chromium Fork for AI Agents Achieves 90.5% on Mind2Web — Show HN 155 PointsABP is a Chromium fork that solves AI browser automation timing bugs by freezing JavaScript execution and rendering after each agent action, capturing the resulting DOM state and notable events, then returning a screenshot and structured summary before unfreezing. This eliminates the stale-DOM race conditions that cause modal-timing failures, autocomplete interference, and download-tracking errors that plague Playwright-based agents. It benchmarks at 90.5% on Online Mind2Web (Claude Opus 4.6) including 85.51% on hard tasks; installs as an MCP server via npx.

Graph relationships

BENCHMARKED AGAINST
Agent Browser Protocol
ABP achieves 90.5% on Mind2Web benchmark with Claude Opus 4.6.
Source finding
BENCHMARKED AGAINST
Agent Browser Protocol -> Mind2Web
ABP achieves 90.5% on Mind2Web benchmark with Claude Opus 4.6.
Source finding

Source trail

Microsoft's July 23 release OpenAI spun up First hyperscaler to productionize computer use Overman & Bayati This paper Show HN GitHub arXiv

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…