Alibaba's Qwen3.8-Max Claims Frontier-Rivaling Scores — AI the News That's Fit to Prompt #87
| A curated AI news digest |
| AI the News That's Fit to Prompt |
| MONDAY, AUGUST 3, 2026 · VOL. 1, NO. 87 · 20 STORIES |
| Lead · AI Models |
|
Alibaba's Qwen3.8-Max Claims Benchmark Scores Rivaling Anthropic
A 2.4-trillion-parameter mixture-of-experts model with roughly 95 billion parameters active per request, a million-token context, and benchmark scores Alibaba puts within about two points of the Western frontier. The weights go to Hugging Face next week. Fable 5 was briefly considered capable enough to warrant export controls; a model claiming the same neighborhood, released open-weight out of Hangzhou, is a fact about what export controls can hold. |
| Also in this issue |
|
AI Safety Here's Why AI Agents Lie and Cheat to Reach Their Goals |
|
AI Safety Now There Are Two Labs Whose Sandboxed Models Got Out |
|
Regulation Europeans Are About to Find Out How Entrenched AI Is in Their Daily Lives |
|
Opinion OpenAI's Amazing, but Vastly Oversold, New Model Astra |
|
The Money To Power AI, Khosla and a16z Bet This Startup Can Reinvent Mining |
|
Technology Google Pulls Its Earth AI Image Tool Less Than a Day After Launch |
|
AI & Politics The Reporters at This News Site Are AI Bots |
|
AI & Culture Fender's CEO Seems to Think Your Bandmates Are Just Analog AI |
|
AI & Culture The AI Industry's Biggest Flex Isn't a Model. It's a Cap. |
|
Creators Is Paying Artists Enough to Convince Them to Embrace AI? |
| ...plus 9 more stories, a 10-minute audio digest, a cartoon, and an AI Acronym Quiz. |
| Read the Full Issue |
|
AI the News That's Fit to Prompt The stories that matter, minus the hype. |
Don't miss what's next. Subscribe to AI the News That's Fit to Prompt: