LATEST AI NEWS
AI NEWS
Anthropic Launches Claude Opus 5 at Half the Cost
AI NEWS
YouTube Adds AI Thumbnail Generation to Studio
HEALTH
AI-Generated ‘Doctors’ Spread Health Myths on TikTok
SOCIAL MEDIA
Midjourney Buys the Astrology App Co-Star
OpenAI Brings Voice Control to the ChatGPT Desktop App
OpenAI has updated its ChatGPT desktop app with ChatGPT Voice, letting users speak to control AI agents and complete multi-step tasks. Powered by the new GPT-Live models, it works with ChatGPT Work and Codex, and on macOS can read what’s on screen.
Source: TechCrunch
AI NEWS
Anthropic Launches Claude Opus 5 at Half the Cost
Anthropic has released Claude Opus 5, a new model it says delivers nearly all the intelligence of its top-tier Claude Fable 5 at half the cost. Available immediately across all of Anthropic’s platforms, Opus 5 becomes the default model on the Claude Max consumer tier and the strongest option on Claude Pro — a launch that signals the AI race shifting from raw capability toward the economics of everyday use.
The pricing tells the story. Opus 5 holds at $5 per million input tokens and $25 per million output tokens, unchanged from its predecessor Opus 4.8, while roughly doubling performance on key agentic benchmarks. The model ships with an adjustable ‘effort’ setting that lets customers trade intelligence for speed and token savings, so buyers can tune cost against quality per task.
On the agentic terminal-coding benchmark Frontier-Bench v0.1, Anthropic reports Opus 5 scored 43.3% — more than double Opus 4.8’s 18.7% and ahead of Fable 5’s 33.7% — at a lower cost per task. The company says it also set new marks on the ARC-AGI 3 reasoning test and the OSWorld 2.0 computer-use benchmark, while candidly noting it still trails a rival model, Mythos 5, on cybersecurity and biology research.
Early enterprise customers emphasized efficiency over raw scores. Harvey said Opus 5 matched Opus 4.8’s maximum-reasoning mode while generating 26% fewer tokens; Zapier’s Wade Foster said it topped the company’s AutomationBench without spending more tokens than prior Claude models. For a company whose customers pay by the token, doing more with fewer is the product itself.
Anthropic, valued at roughly $380 billion in its latest round, also shipped a Fast mode running about 2.5 times default speed, automatic API fallback routing, and prompt-cache retention across mid-conversation tool changes. Developers can call the model as claude-opus-5 today, though when a safety classifier triggers, requests fall back to the less capable Opus 4.8.
Source: VentureBeat
Robi’s Insights:
Holding last year’s price while doubling agentic scores is a stealth price cut Anthropic is buying workloads, not headlines.
The “effort” dial matters more than any benchmark: teams can now meter intelligence against cost the way they meter cloud compute.
Fewer tokens per task is the real pitch; inference cost is a board-line item, and Opus 5 attacks it directly.
Positioning Opus below Fable admits benchmarks miss duration bounded tasks versus long-horizon work is the new dividing line.
Self-verifying behavior compresses human review, the hidden cost that quietly decides whether an AI pilot ever ships.
The classifier fallback to Opus 4.8 means, in practice, your model can be swapped mid-request plan agents accordingly.
Robi’s Remarks:
“Every lab spent three years bragging about its best model on its best day. Anthropic just noticed the money is in a decent model on an ordinary Tuesday priced accordingly.”
OTHER IN AI NEWS
YouTube Brings AI Thumbnail Generation to Studio: YouTube is rolling out AI-generated thumbnails inside Ask Studio for long-form videos, letting creators request and refine a thumbnail by chatting, while Partner Program members can now upload fully custom thumbnails for Shorts and pick from three suggested frames on desktop.
Source: YouTube Blog
SOCIAL MEDIA
Midjourney Buys the Astrology App Co-Star
Midjourney, the AI lab best known for its Discord-based image and video generators, has acquired Co-Star, the social astrology app, according to Bloomberg. The deal terms were not disclosed. Co-Star’s team of roughly two dozen employees is joining Midjourney, bringing the kind of polished consumer-app experience the lab has conspicuously lacked despite its huge user base. Co-Star, which reportedly has about 4.3 million monthly active users,
helps people generate detailed astrological birth charts and compare them with friends inside the app. It leans on a mix of AI and human writers to produce horoscopes, compatibility readings, and daily push-notification advice a proven template for the kind of personalized, habit-forming content Midjourney has never shipped directly.
The purchase fits a company casting an unusually wide net. Midjourney is still synonymous with the Discord server where users type prompts to make images, and it has no standalone consumer app of its own. Yet it is simultaneously building a medical arm and other side ventures, so bolting on an astrology brand looks less like a tidy strategy than a lab following its curiosity and its cash.
For Midjourney, the real prize may be less the star charts than the people who made them. A firm that generates images inside a chat server has never had to design, ship, and retain users on an app of its own. The crew behind one of astrology’s stickiest products could finally give it that muscle.
Source: TechCrunch
🤖 Robi’s Take :
“An image lab with no app of its own buying the horoscope people is peak 2026 when your roadmap is ‘vibes,’ acquiring a company that monetizes them is almost rational.”
OTHER IN SOCIALS
Instagram Cracks Down on Meta Glasses Harassment: Instagram head Adam Mosseri says the platform will now remove videos filmed on Meta glasses that harass strangers including ‘pickup line’ clips of men approaching women in public and at least two pickup-artist accounts, each once topping a million followers, have already been deactivated.
Source: Business Insider
HEALTH
AI-Generated ‘Doctors’ Are Spreading Health Myths on TikTok
AI-generated ‘doctors’ are racking up millions of views on TikTok while dispensing dubious medical advice, according to research that experts warn poses a ‘huge danger to public safety.’ Deepfake accounts built to resemble real physicians are peddling medical myths and so-called miracle cures.
The study, published this month by the digital marketing agency Hallam, found AI-generated content in 40% of top health-related TikTok videos. For some search terms, such as ‘health tips,’ 84% of clips were classed as AI-generated or AI-assisted, and the top AI doctor videos averaged 2.5 million views each.
The avatars pushed disproven cancer claims that microwaving food in plastic, using deodorant, or sleeping beside your phone causes cancer, all refuted by Cancer Research UK alongside invented remedies and a fictional anti-ageing product researchers could find no evidence exists.
Officials lined up to sound the alarm. Britain’s NHS England medical director and the Royal College of GPs both flagged a real risk of harm, while a UCL misinformation researcher described the trend as the ‘industrialised exploitation of trust.’
TikTok disputed the findings as coming from a firm with commercial interests, noting it removes harmful health misinformation and partners with the WHO and NHS. But with roughly one in five people already turning to social media for health information, the volume is the point.
Source: The Guardian
🤖 Robi’s Take :
“A recommendation engine can’t tell a physician from a prompt in a lab coat — it just optimizes for watch time. ‘Industrialised exploitation of trust’ is the platform working exactly as designed.”
DAILY AI TOOL
AI Tool You Did Not Know You Needed
Problem: You need a polished deck by tomorrow morning, but staring at a blank first slide somehow eats the entire evening.
AI Tool: Gamma turns a short text prompt or rough outline into a fully formatted presentation, document, or shareable webpage you can edit.
Solution: Draft your argument in plain text, let Gamma handle the design, and spend your time refining the story instead of nudging boxes around.
PROMPT OF THE DAY
Ship an AI Cost Policy
Prompt: You are a FinOps strategist specializing in enterprise AI adoption. Your task is to design a model-spend governance plan for a mid-sized software company.
Your framework should include: (1) a tiered model-routing policy by task, (2) per-team token budgets, (3) caching and effort-setting defaults, (4) a human-review threshold for agent output, (5) a monthly cost-and-quality review ritual, and (6) success metrics such as cost per resolved task, tokens per workflow, and a target 30% inference-spend reduction in two quarters. Align every choice to measurable business outcomes, not model hype.
SPOT THE FAKE
Can you outsmart AI?
We’ve got a visual challenge for you: one of the two images below is 100% real, the other is crafted by AI.
Click option below A or B. 👇
A
B
BEFORE YOU GO
Ready to take your AI journey further?
For the latest AI updates and insights Click Here 👉 Explore our Weekly Newsletter
For in-depth AI tutorials and updates 👉 Subscribe to our YouTube Channel
AT THE END
Craving more AI chaos?
That's it for today!
Read Daily AI News at BitBiased.AI. Support us by following us on LinkedIn and X ( Twitter ).
Thanks for reading -Stay Curious and a Bit Biased for AI – Robi & the BitBiased.AI team
