OpenAI Launches GPT-6 Sol and Luna, Cuts API Prices by 50%

Key Takeaways

• OpenAI cut API prices for GPT-6 Sol and Luna by 50% compared to GPT-5.6, with Sol making about half as many factual errors as its predecessor.
• Anthropic released Opus 5.5, offering improved writing clarity and lower cost per token for coding and complex knowledge tasks.
• Fastly reported AI traffic grew 6.5 times faster than human traffic from January through May, with machine-generated traffic exceeding 50% of network traffic.
• Meta's Muse AI assistant reached the top of Apple's U.S. App Store shortly after its September 8 launch, helping drive an 11.4% share price increase.
• Applied Materials announced a $5 billion, decade-long plan for a semiconductor research park in India, with shares up 4.44% on the session.
• Darktrace's SECURE AI integrates with AWS, Anthropic, Microsoft and OpenAI, scanning over 80% of monitored customers' generative AI usage for risks.
• Prosus appointed a new AI leadership team including a Chief AI Officer, CTO and Chief Product Officer to build products from one billion customer data points.
• Okta is positioning itself as a gatekeeper for AI agents, expanding tools to detect and prevent misuse in enterprise settings.
• EASA has delayed its planned 2025 certification of a Level 1 AI system in aviation, citing the need for more research and testing.
• Kambi's AI platform processed over 100 million bets during the 2026 FIFA World Cup with zero downtime, now trading more than 70% of all bets.

Anthropic and OpenAI release cheaper and more powerful AI models

Anthropic and OpenAI both announced new AI models that are more powerful and cost less to use. Anthropic released Opus 5.5, which writes more clearly and disobeys rules less often while costing less per token. OpenAI released GPT-6 Sol and GPT-6 Luna, which are better at facts and coding. The new models are available through various platforms and subscription tiers.

OpenAI GPT-6 Sol doubles accuracy for half the cost

OpenAI released GPT-6 Sol and GPT-6 Luna, reporting major accuracy gains and lower costs compared to GPT-5.6. GPT-6 Sol makes about half as many factual errors and costs less per task than Anthropic's Claude Opus 5. The models also communicate more clearly and use prompt caching to cut costs by up to 90 percent. OpenAI says the models show improved alignment and fewer attempts to bypass safety rules.

OpenAI introduces GPT-6 Sol and Luna models

OpenAI introduced GPT-6 Sol and GPT-6 Luna as more affordable models that bring advanced AI to everyday tasks. The models use similar training methods as GPT-6 Astra, bringing strong performance in coding, factuality and alignment to lower price points. OpenAI cut API prices by 50 percent compared to GPT-5.6 pricing. The company says these improvements make advanced AI practical for more users at scale.

OpenAI launches GPT-6 Sol and Luna with lower cost and fewer errors

OpenAI launched GPT-6 Sol and GPT-6 Luna, highlighting lower costs and improved accuracy. The models are available at half the cost of the GPT-5.6 series due to improvements in caching and inference. OpenAI says GPT-6 Sol makes about half as many mistakes as its predecessor on factuality tests. The models are now available in ChatGPT Work, Codex, the ChatGPT API and the desktop app for some users.

Anthropic and OpenAI models offer more performance for less money

Anthropic and OpenAI both released new AI models focused on lowering costs while keeping strong performance. Anthropic announced Opus 5.5 for coding and complex knowledge tasks. OpenAI announced GPT-6 Sol and Luna for efficiency and speed. The new models promise similar performance to earlier versions but at a lower cost, making them more accessible to a wider range of users.

OpenAI launches GPT-6 Sol and Luna at half the price of GPT-5.6

OpenAI added GPT-6 Sol and GPT-6 Luna to its GPT-6 lineup below the top GPT-6 Astra model. Both models carry forward improvements in coding, accuracy, and professional work from Astra. API prices for Sol and Luna are 50% lower than GPT-5.6 pricing, with input and output costs cut across the board. OpenAI says the price drop comes from better caching and inference efficiency passed on to developers.

GPT-6 Sol shows small intelligence gains over GPT-5.6 Sol at lower cost

Independent firm Artificial Analysis tested GPT-6 Sol and GPT-6 Luna against their GPT-5.6 predecessors. GPT-6 Sol scored 48 on the Intelligence Index versus GPT-5.6 Sol's 47, while GPT-6 Luna scored 37, tied with its predecessor. Running costs dropped about 50% for Sol and 60% for Luna per task. However, GPT-6 Luna scored slightly lower on the Coding Agent Index than GPT-5.6 Luna, despite being cheaper.

Fastly unveils AI security tools ahead of Investor Day

Fastly introduced AI Runtime Control, AI Firewall, and new API Security features to help organizations manage AI systems. The company reported that machine-generated traffic made up over 50% of its network traffic in July and August. AI traffic grew 6.5 times faster than human traffic from January through May. These announcements add momentum heading into Fastly's Investor Day on Tuesday.

Fastly Investor Day focuses on growth and AI security push

Fastly's Investor Day highlights management's strategy, product roadmap, and financial outlook. Investors are watching for updates on revenue growth, margins, and demand in security and edge cloud. Fastly's next earnings report is set for November 4, 2026, with estimated EPS of 8 cents and revenue of $187.59 million. Analysts from RBC Capital, Piper Sandler, and Keybanc recently updated their ratings and price targets.

Fastly stock rises as AI security launch precedes Investor Day

Fastly shares traded at $27.75 in premarket, up 1.20%, ahead of Investor Day. The stock sits 19.7% above its 20-day SMA and 43.2% above its 200-day SMA, with resistance near $30.50 and support at $22.50. Fastly's AI traffic grew 6.5 times faster than human traffic from January through May. The company launched new AI security tools including AI Runtime Control and AI Firewall.

Meta shares jump 11.4% as Muse AI assistant gains traction

Meta Platforms shares rose 11.4% after Wells Fargo raised its price target to $796 from $640. The company's Muse AI assistant launched on September 8 and reached the top spot on Apple's U.S. App Store. Muse can handle research, form completion and online purchases. The rally came ahead of Meta Connect, with investors watching for new AI product updates.

Meta stock climbs on analyst upgrade and AI momentum

Meta stock surged after Wells Fargo lifted its price target while keeping its rating unchanged. The firm cited strong early adoption of the Muse AI assistant and progress across Meta's AI products. Muse reached the top position on Apple's U.S. App Store shortly after its September 8 launch. Meta shares have risen about 35% during September, adding to a broader tech rally where the Nasdaq Composite gained 2.3%.

Chip stocks rebound as Applied Materials announces $5B India plan

Chip stocks recovered after a rough September sell-off, with Applied Materials leading the gains. The company unveiled a $5 billion, decade-long plan to build a semiconductor research park on 140 acres in India. Morgan Stanley raised its price target to $454 from $432, and other analysts also lifted targets. Applied Materials shares were up 4.44% on the session and gained about 64% year-to-date.

Two AI cloud stocks to watch before Anthropic IPO

Anthropic, an AI startup, is expected to go public in November, and investors are watching closely. Two AI cloud companies highlighted as potential winners are Anthropic and Cerebras. Both companies have made advances in artificial intelligence and work with major tech firms like Google and Microsoft. Their AI technology is used in natural language processing, computer vision and robotics. Investors see strong growth potential as the IPO approaches.

Darktrace launches SECURE AI to monitor enterprise AI use

Darktrace made its SECURE AI product generally available to help companies track AI usage across their organizations. Data from about 8,200 deployments showed over 80% of monitored customers used generative AI services in August. The product integrates with Amazon Web Services, Anthropic, Microsoft and OpenAI. It scans sessions in ChatGPT Enterprise, Claude, Microsoft Copilot and Amazon Bedrock for risks like jailbreak attempts and data exposure. SECURE AI is available now as a standalone product or part of the Darktrace Behavioral Defense Platform.

Prosus appoints new AI and technology leaders

Prosus announced a new leadership team for AI and technology. Euro Beinat became Chief AI Officer, Paul van der Boor became Chief Technology Officer, and Nalden joined as Chief Product Officer. The team will build products using an intelligence layer based on data from one billion customers. This work aims to boost AI-led growth for partners and SMEs in the Prosus ecosystem.

Okta expands AI agent controls and ambitions

Okta is expanding how enterprises can find, manage and stop AI agents. Ric Smith, President of Products, calls Okta the gatekeeper for AI agents today. The company wants to move beyond security into customer and employee experience. New features will help detect and prevent AI agent misuse while staying flexible for different business needs.

EASA keeps cautious approach to AI certification

The European Union Aviation Safety Agency has taken a slow approach to AI. It had planned to certify a Level 1 AI system by 2025 but has not done so yet. Experts disagree on how fast AI should be adopted in aviation. EASA says a careful approach is best because the technology still needs more research and testing.

Kambi uses AI for first fully AI-powered World Cup

Kambi's AI trading platform was tested at the 2026 FIFA World Cup. Over 100 million bets were placed with zero downtime and an 18% operator trading margin. AI now trades more than 70% of all bets across Kambi's network. The tournament showed a big rise in personalized bets, with Bet Builder bets doubling compared to 2022.

Sources

• Anthropic And OpenAI Announce More Powerful (And Cheaper) AI Models
• OpenAI's GPT-6 Sol doubles its accuracy rate - for half the cost
• Introducing GPT-6 Sol and Luna
• OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes
• New Anthropic, OpenAI models make same promise: A little more for a lot less money
• OpenAI Releases GPT 6 Sol And GPT 6 Luna With 50% Lower Prices Than GPT 5.6
• GPT 6 Sol Shows Modest Gain Over GPT 5.6 Sol On Artificial Analysis Intelligence Index, But At A Much Cheaper Price
• Fastly’s AI Security Push Adds Heat Ahead Of Investor Day
• Fastly’s AI Security Push Adds Heat Ahead Of Investor Day
• Fastly’s AI Security Push Adds Heat Ahead Of Investor Day
• Meta Stock Soars After Jaw-Dropping Price Target Hike on AI Products
• Meta Stock Soars After Jaw-Dropping Price Target Hike on AI Products
• Chip stocks snap back on multi-billion-dollar AI bet - AI Strategy got in early
• Got $5,000?: 2 Magnificent AI Cloud Stocks to Buy Before the Anthropic IPO
• Exclusive: Darktrace makes SECURE AI generally available to monitor enterprise AI use
• Prosus names AI and technology leaders building the next chapter
• ‘The gatekeeper’ - Okta expands AI agent controls, with ambitions beyond security?
• EASA Carries On With Cautious Approach To Artificial Intelligence
• Trading transformed: Inside Kambi’s first fully AI-powered World Cup