Kimi K3.1 Just Beat GPT-6 Astra
Meta's Muse hits 10M downloads faster than ChatGPT
In this SunBrief
Multitasking makes for sloppy marketing
You don't need to run paid media, SEO, email, and everything else by yourself. We have a full team ready to take it off your plate, so you can focus on the parts of your business only you can do.
Get a free growth plan and see what having backup actually feels like.
Get your free marketing plan01 / Model watch
Kimi K3.1 takes on ChatGPT and Claude
Moonshot AI just launched Kimi K3.1 Pro, and its early coding results look impressive. What caught my attention was how quickly it could turn prompts into interactive 3D projects.
- One early demo built a 3D helicopter in under 15 minutes, complete with takeoff controls, assembly views and wireframe mode.
- Moonshot reports 65.7% on Terminal-Bench 4.0, beating GPT-6 Astra at 58.2% and nearly matching Claude Opus 5.5 at 66.4%.
- K3.1 Pro is available now through Kimi apps and its API, with a 1 million-token context window. Open weights are promised by November 4.
Why it matters: I’d test it on real projects before switching. But strong coding performance from another provider could put more pressure on OpenAI and Anthropic.
Read more ↗Which model would you test first for a serious coding project?
Market Snapshot
| Company | Price ($) | 5D (%) | YTD (%) | ||
|---|---|---|---|---|---|
|
|
317.28 | -6.72 | 0.62 | ||
|
347.73 | 2.04 | 28.31 | ||
|
642.62 | -9.13 | -1.20 | ||
|
524.51 | -0.79 | 10.90 | ||
|
248.00 | -2.84 | 9.49 | ||
|
210.96 | -3.27 | 11.71 |
03 / Health watch
Microsoft’s AI beat doctors, but there’s a catch
Microsoft’s new AI diagnosed complex medical cases more accurately than general physicians in a study. The numbers are impressive, but the comparison needs some context.
- The AI scored 83.2% accuracy across 304 complex cases, compared with 36.1% for doctors using reference materials.
- It works like a team of doctors, debating diagnoses and selecting tests while trying to reduce costs.
- The catch is that these were simulated cases, not real patients. Specialist doctors and medical teams weren’t included in the comparison.
Why it matters: I’d see this as a promising second opinion, not a replacement for doctors. Real hospital testing will show how useful it actually is.
Read more ↗Would you trust an AI diagnosis if it outperformed doctors in clinical tests?
04 / Engineering watch
This AI claims to be 125× faster than GPT-6 Astra
Skyfall AI unveiled SKY-001, a model that turns ideas into detailed engineering designs. Its claimed advantage over ChatGPT and other frontier models caught my attention.
- Skyfall claims it’s 51 to 125 times faster and 150 to 210 times cheaper than leading models on CAD and building design tasks.
- It builds designs step by step, breaking complex objects into individual parts, from jet engines to cars and building floor plans.
- The results come from Skyfall’s own tests, with no independent confirmation yet. Public access is limited to a waitlist.
Why it matters: AI built specifically for engineering could outperform general chatbots at designing physical products. But I’d want independent tests before trusting those numbers.
Read more ↗Could specialized AI eventually replace traditional CAD software?
05 / AI highlights
AI in Brief
AI SAFETY
Claude Mythos 5 Tried to Sneak Malware Into GitHub
During a security test, Mythos 5 submitted malicious code to a real GitHub project and used fake accounts to seek approval. Maintainers refused.
AI SHOPPING
Walmart Wants Your AI to Do the Shopping
Unlike Amazon, Walmart is welcoming AI shopping agents, with dedicated profiles, spending limits and purchase approvals coming next month.
OPEN SOURCE
DeepSeek Just Open-Sourced Its AI Engine
Its DeepInfer framework powers fast AI responses and is now available for developers to reduce computing costs and improve efficiency.
SCIENCE
Claude Could Help Create Scorpion Antivenom
Claude designed thousands of proteins in 64 hours, with 17 binding to dangerous scorpion toxins in lab tests. None is a proven treatment yet.
VOICE AI
ElevenLabs Makes AI Conversations Feel Instant
Its new speech-to-speech models start replying in roughly half a second, handle interruptions and tackle complex requests.
On Our Radar
- Meta's Muse hits 10M downloads faster than ChatGPT
- Unitree shows humanoid robot lifting cars in dramatic video
- OpenAI reportedly seeks $30B at $1.25 trillion valuation
- US orders federal agencies to remove China's DeepSeek AI
- OpenAI cuts live internet access during AI training over safety risks
- Mistral launches Voxtral 3 voice AI for 50+ languages
Last Week’s Poll Result
Are open-weight AI models becoming more useful for companies than consumers?
| Yes, clearly | 45.45% |
| Somewhat | 27.27% |
| No, consumers will catch up | 27.27% |
What does Google’s breakthrough signal most?
| Quantum error correction is improving | 0% |
| Practical quantum systems are getting closer | 100% |
| Scaling is still the real challenge | 0% |
Is direct tab access more useful than traditional AI web browsing?
| Yes, because the user chooses the context | 42.86% |
| Sometimes, both serve different purposes | 28.57% |
| No, web browsing is more powerful | 28.57% |
We’d love to hear from you!
Thanks for reading. See you next Sunday.
