Wednesday 23 September 2026
Top stories today:
Claude Opus 5.5 debuts as Anthropic’s 1st model since Dario’s call
OpenAI’s GPT-6 Sol, Luna debut with gains at 50% lower cost
Apple to develop screenless health tracker similar to Whoop
Micro1 to raise $100M+ at $4B valuation as ARR tops $500M
Binance acquires $100M Circle stake, signs 5-year deal for USDC
0. Data & Calendar
All values as of 6 AM ET / 3 AM PT, other than S&P500 and NASDAQ close (4 PM ET / 1 PM PT).
All times are ET.
Listen to our AI-generated podcast summarizing today’s newsletter (beware of hallucinations):
1. Claude Opus 5.5 debuts as Anthropic’s first model since Dario Amodei’s call to slow AI development
Claude Opus 5.5 topped the Artificial Analysis Intelligence Index, surpassing Fable 5.1:
Costs $4/$20 per 1M input/output tokens, -20% from Opus 5, vs. GPT-6 Astra’s $10/$50 and Muse Spark 1.3’s $1.25/$4.25.
Features enhanced safeguards to combat risky behavior, Anthropic said.
Anthropic has also raised its 5-hour usage limits by 20% on Pro, Max, and Team plans and introduced a rate-limit reset for subscribers.
Hear real-world success stories from Vercel and Notion on automating workflows, speeding up support, and boosting sales productivity with AI agents in Slack.
Learn how to identify the right AI agent opportunities for your team and measure ROI and adoption.
Get integration strategies that scale, straight from partners already doing it.
*Sponsored.
3. OpenAI’s GPT-6 Sol, Luna debut with performance gains at 50% lower cost than GPT-5.6
GPT-6 Sol scored 48 on the Artificial Analysis Intelligence Index, while GPT-6 Luna’s score remained unchanged:
GPT-6 Sol costs $2/$10 per 1M input/output tokens, vs. Muse Spark 1.3’s $1.25/$4.25, Grok 4.7’s $2/$6, and MiMo-V2.6 Pro’s $0.435/$0.87.
GPT-6 Luna costs $0.10/$0.50, vs. Gemini 3.8 Flash’s $0.75/$3.75 and DeepSeek V4.1 Flash’s $0.006/$1.20 (peak rate).
GPT-6 Sol makes 50% fewer mistakes than GPT-5.6 Sol, OpenAI said.
GPT-6 Sol and Luna both hallucinate less than their predecessors at max effort
Separately, OpenAI plans third-party safety evaluations across its AI model development and deployment.
Training AI on finance requires experts who can evaluate models on valuation, portfolio theory, and risk modeling—not just general annotators.
Athyna Intelligence delivers financial analysts, economists, and quant researchers from Latin America for RLHF, evals, and reasoning tasks.
Same US time zone. Vetted in days. 40–60% savings.
*Sponsored.
5. Apple said to develop screenless health tracker similar to Whoop
Apple is now reportedly developing prototypes after exploring the category for several months.
The project remains in a preliminary phase, and the device likely won’t debut before 2028 if Apple proceeds with a release.
Garmin shares -2.5% to a session low on the news but pared losses to close Tue. +2%, while Apple rose as much as 1% before closing +0.23%.
Get the chance to peek inside founders and leaders’ brains and see how they think about going from zero to 1 and beyond.
Join thousands of weekly readers at Google, OpenAI, Stripe, TikTok, Sequoia, and more.
Check all the tools and more here, and outperform the competition.
*Sponsored.
7. Micro1 said to raise $100M+ at $4B valuation as ARR tops $500M
Two xAI co-founders are reportedly participating in the round.
Microsoft, Amazon, and 1X are among its customers.
