Anthropic’s Claude Haiku 5.5 debuts with controls to adjust reasoning effort

retained text

Thursday 8 October 2026
Top stories today:
- Anthropic’s Claude Haiku 5.5 debuts with effort controls
- OpenAI’s new math work criticized as power, not scholarship
- ChatGPT gets GPT-6 with interactive Intelligent UI
- Google launches AI-powered no-code game platform Playground
- Isomorphic Labs seeks funding at up to $50B valuation
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).
No major events today.
Listen to our AI-generated podcast summarizing today’s newsletter (beware of hallucinations):
1. Anthropic’s Claude Haiku 5.5 debuts with controls to adjust reasoning effort
Claude Haiku 5.5 scored 43 on the Artificial Analysis Intelligence Index, comparable with Kimi K3 and ahead of Gemini 3.8 Flash:
- Costs $0.10/$0.50 per 1M input/output tokens (<100K tokens) or $0.50/$2.50 (>100K tokens), vs. Kimi K3’s $3/$15, Gemini 3.8 Flash’s $0.75/$3.75, and GLM-5.3-Flash’s $0.15/$0.50.
- Costs -75% to run vs. Haiku 4.5, Anthropic said.
- Anthropic also cut Sonnet 5.5’s cache-read price to $0.10 from $0.20 and added monthly API credits.
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.
3. OpenAI’s new math work shows power, not scholarship, mathematicians’ group says
- “Mathematicians did not ask for this work to be done,” the Association for Human Mathematics wrote.
- Tue.: OpenAI released 719 manuscripts detailing findings on 377 math problems by its internal model.
- Relatedly, Vitalik Buterin warned of AI-driven cryptography risks, backing blockchain “bunker mode.”
- Separately, 3 fired OpenAI researchers urged AI labs to halt work that could impair AI monitoring, saying their firings have a “chilling” effect.
Athyna Intelligence builds what comes after.
PhD researchers and domain experts author verified, containerized tasks in the Terminal-Bench and GDPval mold, spanning coding, finance, accounting, engineering, and more.
On our harder tasks, frontier models score 0 out of 5.
*Sponsored.
5. ChatGPT gets GPT-6 with interactive Intelligent UI featuring charts, buttons in responses
- GPT-6 enables ChatGPT to answer while continuing to think in the background.
- Rolling out globally to ChatGPT Plus, Pro, Business, and Enterprise users.
- GPT-6 Sol powers Plus, Pro, Business, and Enterprise tiers, while GPT-6 Luna powers Free and Go.
Roko's Basilisk is where engineers, founders, and future-thinking people get their daily dose of tech and AI news. And they read it.
The newsletter averages a 41% open rate and a 5% click-through rate, with 98% of readers based in the US. Featured placements sit at the very top of the email, with your image, copy, and CTA button front and center.
Want to reach the people building what's next? Let's talk.
*Sponsored.
7. Google launches AI-powered no-code game platform Playground for U.S. users 18+
- Gemini, Nano Banana, and Lyria power the platform.
- Users can create games using AI prompts.
- Google separately launched SynthID Detector to identify AI-generated media across 20 file formats

document fetched 2026-10-24 10:03 · canonical https://bayareatimes.com/p/anthropic-s-claude-haiku-5-5-debuts-with-controls-to-adjust-reasoning-effort