This is actually a pretty surprising and something that should lead companies to change how they are thinking about hosting. Model performance for the open weights GPT model vary by meaningful amounts depending on who is hosting it, with Azure & AWS being low. Worth watching.
GENERATIVE AI
-
Genie 3 and Gemini 2.5 Advance Toward AGI with Better Benchmarks
By
–
To reach artificial general intelligence, we’ll need both advanced AI models that can think and understand the world around us and better benchmarks to evaluate their progress.
— Google AI (@GoogleAI) 12 août 2025
Listen in as @demishassabis and @OfficialLoganK chat about how our new world model Genie 3, Gemini 2.5… pic.twitter.com/aYSdmCqzyKTo reach artificial general intelligence, we’ll need both advanced AI models that can think and understand the world around us and better benchmarks to evaluate their progress. Listen in as @demishassabis and @OfficialLoganK chat about how our new world model Genie 3, Gemini 2.5
-
Comparing Kimi K2 and GPT-OSS Integration in Development Workflows
By
–
For those who’ve built with both Kimi K2 and gpt-oss, how have you integrated them into your workflows and which is working better for you?
-
Claude Opus 4.1 Interface Limitations and Performance Issues
By
–
Every time I try Opus 4.1 in the Claude client, it just breaks – goes around in circles even for relatively simple tasks and then throws up an error. I'm sure it is not that Opus is bad, but it seems to have outgrown its Claude interface shackles
-

Mistral Medium 3.1 Released
By
–

Mistral released Mistral Medium 3.1! It is now available as a default model on Le Chat as well as in the API as mistral-medium-2508
-

Fixing Transformers GPT-OSS MoE Finetuning Issues
By
–
"i thought the transformers gpt-oss MoE finetuning was broken, how did you get it working?"
-
Monitoring and measuring brand performance in the era of AI search
By
–
77% of searches now end with AI-generated answers instead of website clicks. It’s killing traditional SEO. Profound is the first platform that lets brands monitor, measure, and dominate this new AI search world. Here’s how it works
-

Multiple AIs Debating Health Diagnostics More Accurately and Cheaply
By
–
Don’t be tricked by the headlines. This article about AI and health diagnostics is NOT ABOUT HEALTH. This is about MULTIPLE AIs all arguing together, with internet access, more cheaply and at higher accuracy than one expensive expert. The paper is flawed (the human should have
-

Global Leaders Deploy Agentic AI Agents into Production with North
By
–
Global leaders are putting agentic AI into production with North. Integrated AI agents Real productivity gains Private, secure deployments See how RBC, LG CNS, and Dell are scaling with North: http://
cohere.com/blog/north-ga -

Claude Sonnet 4 1M Context Window Access Requirements
By
–
Notes on the new 1m context window for Claude Sonnet 4: https://
simonwillison.net/2025/Aug/12/cl
aude-sonnet-4-1m/
… You need to send a beta header of context-1m-2025-08-07 and be on tier 4, which means you have purchased at least $400 in API credits
