The changes are focused on making ChatGPT more direct, factually accurate, and consistent across quick questions and deeper reasoning tasks.
OpenAI rolls out a major ChatGPT upgrade, even if you don’t pay for it
Source: BleepingComputer
As you can see in the screenshot, you can slide GPT's reasoning from Instant to High. Instant would respond almost immediately, while High reasoning could take several minutes.
You can keep the slider low for everyday questions or increase the reasoning effort for tasks such as research, coding, writing, planning, and complex decisions.
GPT-5.6 Sol is getting more focused and reliable answers
OpenAI says the updated GPT-5.6 Sol has been tuned specifically for how people use ChatGPT, rather than for the longer-running agentic workflows found in Codex and ChatGPT Work.
"It delivers more focused answers, adapts its level of detail to the question, avoids unnecessary formatting, and offers a helpful correction when simply agreeing wouldn't be useful," OpenAI said.
For example, if you ask simple questions, GPT should answer directly without adding unnecessary background. For more complicated work, it should provide enough detail while keeping the main recommendation easy to identify.
A question like "Who is the president of the United States?" will have a single-line answer, but a question like "How are elections held in the US?" will have a longer answer with background information.
OpenAI also claims that GPT-5.6 Sol is less likely to make factual mistakes when handling dates, numbers, sources, legal rules, medical information, financial questions, or assumptions.
In internal tests, OpenAI found that responses containing at least one factual error were 68% less common with GPT-5.6 Sol than with GPT-5.5 Instant.
Moreover, GPT-5.6 Luna reduced such responses by 62% in the same evaluation.
Free users are getting unlimited GPT-5.6 Luna chats
GPT-5.6 Luna will become the default model for Free and Go users this week, and some users are already seeing the change. I spotted it in my OpenAI account with a Go subscription that costs approximately $10.
Source: OpenAI
Starting next week, OpenAI also plans to remove the rate limit for text chats on those plans, although usage will remain subject to abuse protections. However, limits will still apply to file uploads, image generation, and other ChatGPT tools.
With a free account, you will still receive a new Think button that gives GPT-5.6 Luna additional time to reason through harder questions.
"For questions that require deeper reasoning, Free users can tap the new Think button to give GPT-5.6 Luna more time to work through the answer," OpenAI noted.
The update does not affect GPT-5.6 Sol in Codex or Work
The revised version of GPT-5.6 Sol is designed for everyday conversations and will only be available through ChatGPT's regular Chat experience.
OpenAI is also introducing additional protections for users believed to be under 18.
These include stricter boundaries around romantic roleplay, sexual content, dangerous activities, eating disorders, body-image risks, age-restricted goods, and graphic violence.
Today's changes are rolling out gradually.
LATEST ARTICLES
ClickFix attack pushes macOS infostealer for crypto theft attacks
A Go-based malware delivered in ClickFix attacks targeting macOS users is stealing cryptocurrency assets, browser-stored passwords, Apple Keychain data, and cached credentials.
- August 06, 2026
- 06:37 PM
0
Webinar: The state of shadow AI in 2026 (and how attackers are taking advantage)
AI adoption has exploded, but every new app, integration and extension introduces new threats and risks.
The latest webinar from Push Security uncovers the scale of Shadow AI in the enterprise, how attackers are taking advantage, and what security teams can do about it.
Hedge fund cyberattacks tied to BlackFile-linked UNC6671 extortion group
A recent wave of cyberattacks targeting hedge funds, private-equity firms, and other financial organizations has been linked to UNC6671, an extortion group reportedly associated with the BlackFile threat actors.
- August 06, 2026
- 04:07 PM
- 0
Swiss government SharePoint breach compromised 200 accounts
Switzerland's federal IT office says hackers exploited vulnerabilities to breach its Microsoft SharePoint servers and compromised approximately 200 accounts.
- August 06, 2026
- 02:14 PM
- 0
Train for all eight CISSP domains from home for just $20
This CISSP Security & Risk Management Training Bundle is an eight-course bundle for just $19.99 (MSRP: $424), covering every domain included in the Certified Information Systems Security Professional (CISSP) framework exam.
- August 06, 2026
- 02:10 PM
- 0
New TONTOU CPU attack bypasses Spectre v2 fixes, leaks Linux password hashes
Researchers found a way to bypass recent mitigations for Spectre v2 speculative execution side-channel attacks and developed an exploit to leak secrets from Linux machines.
- August 06, 2026
- 02:03 PM
- 0
========================================================
Meta AI model hacked a company during misconfigured cyber test
Meta has become the latest AI company to confirm that one of its models hacked a real organization during cybersecurity testing, as similar incidents continue to emerge following OpenAI'sOpenAI's initial disclosure that its agents breached Hugging Face.
- August 06, 2026
- 12:11 PM
- 0
How AI Exposed a Browser Security Gap that Enterprises Cannot Ignore
AI did not create a new browser security problem. It exposed one that enterprises have long been able to ignore. Skyhigh Security explains why browsers have become a critical control point for governing data movement, AI interactions, and modern work.
- August 06, 2026
- 10:02 AM
- 0
Enhance privacy on up to 10 devices with this 5-year AdGuard VPN deal
Online privacy shouldn't come with a steep price tag, and this AdGuard VPN subscription proves it. For a limited time, new users can lock in five years of protection for $39.97, down from $359.40, which works out to less than a dollar a month.
- August 06, 2026
- 07:09 AM
- 0
Ransom Cartel ransomware creator sentenced to 16 years in prison
Maksim Silnikau, the creator and administrator of the Ransom Cartel ransomware operation, was sentenced to 16 years in prison for his role in ransomware attacks against at least 18 companies worldwide.
- August 05, 2026
- 07:00 PM
- 1
Canadian pleads guilty to Snowflake cloud data-theft attacks
A Canadian man pleaded guilty today to his role in accessing company accounts at cloud storage provider Snowflake and stealing data from at least 165 organizations in a scheme to extort millions of dollars from victims.
- August 05, 2026
- 05:53 PM
- 0
Hackers run khunt post-exploitation toolkit from Oracle database
Hackers exploited a SQL injection vulnerability to install a post-exploitation toolkit directly inside an Oracle database that was used to breach a corporate network.
- August 05, 2026
- 03:55 PM
- 0
This 3-year Surfshark VPN deal covers unlimited devices for $84
Right now, you can get a 3-year Surfshark VPN Starter subscription for $83.99 (MSRP $430), making long-term online protection much more affordable.
- August 05, 2026
- 02:11 PM
- 0
COLDCARD security audit phishing attack installs remote access tool
A phishing campaign is exploiting fears surrounding the recently disclosed COLDCARD wallet vulnerability and suspected $88.6 million Bitcoin theft to trick users into installing ScreenConnect remote access software.
- August 05, 2026
- 01:49 PM
- 0
CISA warns of hackers exploiting Langflow, N-central, Apache Tomcat flaws
The U.S. Cybersecurity and Infrastructure Security Agency is giving federal agencies three days to mitigate vulnerabilities in IBM Langflow, N-central, and Apache Tomcat, all actively exploited.
- August 05, 2026
- 11:51 AM
- 1
Google Blogger locks hundreds of blogs in malware false positive
Google has locked hundreds of Blogger websites after a false positive claimed they violated its "Malware and Similar Malicious Content" policy, with some sites deleted from the platform.
- August 05, 2026
- 10:59 AM
- 1
How AI-powered phishing killed blocklists for good
AI is helping attackers create disposable phishing infrastructure and rapidly evolving toolkits that blocklists cannot track fast enough. Push Security explains why browser-level, technique-based detection offers a more durable defense than relying on domains, signatures, and other known-bad indicators.
- August 05, 2026
- 10:01 AM
- 0
Launch your IT career for $39.99 with this lifetime training bundle
For just $39.99 (MSRP $2,748.75), you'll get lifetime access to the Complete CompTIA & IT Exam Training Bundle, packing together five learning platforms, more than 180 hours of training, and certification prep for some of the industry's most in-demand technologies.
- August 05, 2026
- 07:12 AM
- 0
OpenAI, Anthropic AI agents targeted real people and systems in cyber tests
OpenAI and Anthropic have confirmed that their AI models were involved in separate, newly disclosed third-party cybersecurity testing incidents that resulted in a real website being breached and social engineering attacks against people outside the intended testing boundaries.
- August 04, 2026
- 07:39 PM
- 0
TP-Link patches Omada ZTP flaws allowing hackers to breach networks
TP-Link has patched 15 vulnerabilities in the zero-touch provisioning (ZTP) mechanism of its Omada network devices that could be chained with previously disclosed flaws to achieve remo
