LongCat-2.0: The 1.6T Coding Model That Hid on OpenRouter for Two Months
Meituan's LongCat-2.0 spent two months on OpenRouter as "Owl Alpha." Now open source under MIT, it beats GPT-5.5 on SWE-bench Pro at a ...
Microsoft Frontier Company: $2.5B to Fix Enterprise AI Deployment
Microsoft's $2.5B Frontier Company embeds 6,000 engineers in enterprises to fix AI deployment, launched 48 ...
GPT-5.6 Sol, Terra & Luna: What Developers Need to Know Before GA
GPT-5.6 launched June 26 but most developers cannot access it yet. Here is the pricing, ...
Claude Sonnet 5: The Tokenizer Trap Behind the ‘Cost-Neutral’ Launch
Claude Sonnet 5 dropped June 30 with introductory pricing that looks cost-neutral. After September 1, ...
Venice AI Hits $1B: What the Privacy API Means for Devs
Venice AI raised $65M at a $1B valuation with zero data retention and an OpenAI-compatible ...
UN AI for Good Commission: What Developers Need to Know
The UN just launched an AI governance commission with Jensen Huang, Andy Jassy, and Jack Clark. First meeting July 8. Here’s what it ...
Claude Fable 5’s Safety Filter Is Blocking Your Code
Anthropic redeployed Fable 5 with a new cybersecurity classifier that generates false positives on routine ...
8090 Labs Raises $135M: Enterprise AI Coding Platform
Chamath Palihapitiya takes full-time CEO role at 8090 Labs after a $135M raise. What Software ...
Composer 2.5: Cursor’s Model at $0.07/Task vs Opus 4.7’s $4.10
Cursor's Composer 2.5 scores 79.8% on SWE-Bench at $0.07 per task — one point behind ...
nanochat Tutorial: Train Your Own LLM for $100 (2026)
Andrej Karpathy’s nanochat trains a full ChatGPT-style LLM for around $100 on spot GPU instances. ...