The codex-1 model is state of the art on SWE-Bench Verified, we published the numbers on the blog. More importantly, we optimized it to generate code that people actually want to merge, not just code that scores well on benchmarks!
RESEARCH
-
Anthropic Economic Index Launch with Alex Tamkin
By
–
I love what @AnthropicAI has done to create its Economic Index. No one understands it better than @AlexTamkin
. Join us on Monday and bring your questions! -
New Tab Model Enables Cross-File Jumping Capability
By
–
We've trained a new Tab model that can jump across files pic.twitter.com/RQixD0rCGQ
— Cursor (@cursor_ai) 16 mai 2025We've trained a new Tab model that can jump across files
-
Third-party framework flags AI model flaws and bias issues
By
–
Flaws in AI models — from hallucinations to biased training data — often manifest only when systems are live. Experts propose a new framework for third-party reporting to allow external evaluators to identify and flag these issues effectively.
-

European AI Policy Priorities with EU Commission Vice President
By
–
How do global policymakers envision the future of AI? This week, we hosted Executive Vice President of the European Commission @HennaVirkkunen for a fireside chat with HAI Senior Fellow Susan Athey to explore Europe’s key AI policy priorities.
-
AI Tutoring Programs Boost STEM Higher-Order Thinking
By
–
– Programs lasting 4-8 weeks were the most effective
– Learning performance rose the most in problem-based learning
– AI acting as a tutor vs. tool drove more higher-order thinking gains
– AI boosted higher-order thinking for STEM content/courses most over other types -
LLMs Transforming Education Through Diverse Learning Modalities
By
–
Why I’m hopeful the numbers will keep – Some lag in the mix here due to the older studies. LLMs are making huge leaps!
– More modalities (quizzes, podcasts, etc.) to fit learning styles
– EQ gains, esp important for students who may be struggling/discouraged -

AI Boosts Student Critical Thinking Beyond Grade Improvement
By
–
The fear: AI will make students lazy. The reality? It’s making them smarter. A new meta-analysis of 51 studies (!!) shows AI is actually boosting critical thinking, not just grades. My biggest takeaways: 1/
-
ChatGPT Boosts Student Learning Performance and Higher-Order Thinking Skills
By
–
TLDR: Researchers found using ChatGPT had a large positive impact on students’ learning performance (.867 SD) and a moderate positive impact on both learning perception (.456 SD) and higher-order thinking (.457). Some highlights of where/how to apply AI for max impact:
-
Flu strain extinction: A complex systems phenomenon worth attention
By
–
Complex Systems exists in part because I have some special interests and wanted an excuse to say “Hey did anyone else notice that a line of the flu went extinct? That seems like it is worth more attention than it received. News we can use and all that.”