No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark

SemiAnalysis' new GDPx benchmark tests AI agents on simulated knowledge work from chip design to quant trading, with no model exceeding 12 percent; Claude Opus 5.5 leads.

Read the full story at The Decoder ↗

Timeline · 1 report

2026-11-05 16:46 · The Decoder
No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark

More stories

Claude Code drives wave of community-built apps, tools and workflows — DEV Community — AI
AI Coding Agents Shift From Prompt Files to Deterministic Scripts and Sandboxed Execution — DEV Community — AI
Anthropic launches Claude Personal Finance in the US with read-only bank access — Engadget
Meta acquires Conductor, startup behind app for running parallel AI coding agents — Google DeepMind Blog
DeepSeek V4.1 Flash gains traction across local rigs, benchmarks, and serving platforms — Vercel Blog
Perplexity launches Teammates, always-on AI agents for ongoing work — The Verge AI
Dedicated AI voice recorders gain ground over phone apps with local AI and zero-cloud privacy — DEV Community — AI
AWS expands Bedrock AgentCore with web answers, remote Claude Code sessions and agent governance tools — AWS Machine Learning Blog

Get the daily brief of stories like this at 6:30 every morning →