Claude Fable 5.1 is now live on LobeHub, with a massive jump in agentic performance. 🔥 → 52.6% on Terminal Bench Science, more than 2× Fable 5 → 55.8% on Terminal Bench 4.0, up from 42.0% → Stronger across coding, research, computer use, and business workflows → Up to 45% lower cost on highly agentic workloads Built for long running tasks, deep research, and complex agent workflows. Try Claude Fable 5.1 on LobeHub today.

A CHINESE CLAUDE COMPETITOR CALLED GLM 5.2 JUST LAUNCHED AND IT IS CRUSHING FABLE 5 IN SEVERAL AREAS WHILE COSTING A TENTH OF THE PRICE It launched a day ago and people are already switching over. What you get for free every single day: → 1,000,000 credits daily, enough to complete serious tasks without hitting a wall → A powerful model that outperforms earlier Claude models on agent and tool use → Open source with full control and self hosting so nothing is locked behind a paywall What makes it worth paying attention to over Fable 5: → Better agent and tool use than earlier Claude models → Zai coding plans that improve price to performance ratio significantly → Open source gives you full control and the ability to self host → Better for long cursor coding sessions where Fable 5 starts to struggle → Better daily value and quota use so you actually get more done per day This costs 1/10th of what Fable 5 costs and the base tier is completely free.

Odds on a Fed hike Wednesday went 61.5% → 87.5% in three days. October and December haven't moved a basis point in two. Forget the 87.5%. The story is in the two numbers that didn't move. October: hold 61.5%, hike 36.5%. Identical to yesterday — not a decimal of difference. December: hike 57.5%, hold 39.5%. Also identical. Second day running, the market repriced one meeting and left the rest alone. Which says: the market thinks the Fed moves Wednesday, and does not think that changes the path afterward. "This one" and "how many more" are being bet as two separate things. Volume on the September contract went from $6.94M to $18.35M in a day (+164.6%). The four markets on September's inflation print, added together, hold $7,814. That's 2,349x. And the gap keeps widening: 99x on 09-10, 228x on 09-11, 292x yesterday, 2,349x today. Nobody is betting on the number. Everybody is betting on the reaction. Stocks traded it: S&P −0.48%, Nasdaq −0.56%, and the VIX up 7.95% to 17.1. The 10-year yield went the other way, down 1.4bp to 4.961%. Friday was stocks up, VIX down, yields up. Last night was the exact inverse — and no data landed in between. Three gauges flipped in two sessions during a week with no inflation print at all. That isn't data. That's a room waiting on a meeting. Next: the decision lands Wednesday 09-16, 2:00pm ET — about 38 hours out, with a fresh dot plot. At 87.5%, the hike isn't the news. Not hiking would be.

Later today we'll adjust two parameters: Purchase surcharge: 2.5% → 1.5% Buyback allocation: 40/30/30 → 60/20/20 (purchasers/depositors/burn). We'll continue to monitor and adjust small mechanisms like this for the benefit of the protocol

Every AI has a personality 😅 Claude → anxious overachiever GPT-5 → eager intern, hallucinates Grok → edgy contrarian, hot takes Gemini → quiet underdog, underrated DeepSeek → chaotic genius, spins Qwen → silent assassin, cheap & deadly Pick your model like you pick your team

The Jazz have the highest increase in YoY projected preseason win total: Jazz: 18.5 → 36.5 (+18) Spurs: 44.5 → 60.5 (+16) Wizards: 20.5 → 34.5 (+14) Hornets: 27.5 → 39.5 (+12) Celtics: 41.5 → 51.5 (+10) Suns: 30.5 → 40.5 (+10)

Just rode both in Vegas with @DavidMoss Zoox → fully driverless Hyundai Motional Ioniq 5 → still has a supervised safety driver Wild difference on the Strip. Full comparison video.

Gemini 3.7 Flash — the CHEAP model — tripled its agent score (Terminal-bench 3.0: 5.4% → 14.9%) in THREE WEEKS not a new generation. a point release. 21 days. flash went 3.5 → 3.6 → 3.7 in ~3 months while the price got cut in half

The pace of Grok development across SpaceXAI right now is insane In just over 2 months we got: • July 16 — Grok 4.5 • August 12 — Grok 4.6 • September 21 — Grok 4.7 Three major frontier-model generations in barely 9 weeks And each one has pushed Grok further on coding, agentic work, knowledge tasks, long-running reasoning and efficiency SpaceXAI just shipped 4.5 → 4.6 → 4.7 almost back-to-back That iteration speed is absolutely wild