OpenAI on Accelerating Research with Agents
OpenAI has published a blog post on how agents are accelerating research within the company.
They claim to have achieved the goal set last fall, namely: to have an "automated research intern" by September 2026, a system capable of performing clearly defined research tasks, including those that would take a qualified specialist several days.
Additionally, as of mid-August at OpenAI:
- The median researcher spends over $600 a day on agents at API prices. The 90th percentile already exceeds $7000 tokens per day.
- For every human workday, there are approximately 3.1 "agent-workdays" (back in June, the total working time of agents was less than the total human hours).
- The number of experiments per active experimenter and the success rate of agents across several difficulty tiers is actively growing (difficulty is assessed as the time the task would take a human).
In the post, OpenAI also demanded that such progress-sharing practices become mandatory for all companies in the industry. They argue that this is leading to full recursive self-improvement (RSI), and the public should have the ability to track and discuss it.
Why it matters
AnalysisThis approach could significantly enhance research efficiency, allowing scientists to focus on more complex tasks while agents handle routine work. It may also promote greater transparency and collaboration within the industry.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
JSON — Structured Data for Programs, APIs, and AI Agents
A complete hands-on guide to JSON: syntax specifications, serialization via parse and stringify, REST API consumption, JSON Schema validation, and structured outputs for LLMs.
Four Resume Tips from OpenAI’s Head of Applied AI Development
Wulfie Bain, OpenAI’s head of applied AI development, shared four resume rules: keep it short, remove irrelevant details, highlight outcomes, and lead with the strongest achievements.
AI Radar: GPT-6.1 Terra, Gemini 4 Carbon, Mistral Large 4 and OpenAI/Anthropic pricing changes
A roundup of recent AI news: new LLM releases, OpenAI and Anthropic subscription changes, world models, video generation, and research updates.
Codex Skills for Beginners: Automating AI Agent Workflows
A comprehensive hands-on guide to building and executing Codex Skills: SKILL.md anatomy, explicit vs. auto-triggering, skills vs. scripts vs. tools, Skills API, and safety boundaries.