The internet, freshly baked.

SpaceX snaps up AI coding startup Cursor for $60 billion days after record IPO

SpaceX has agreed to acquire Cursor, an AI-powered coding platform created by Anysphere, for $60 billion in a stock deal, just days after the company's record-breaking initial public offering on the Nasdaq stock exchange. The acquisition, confirmed by Elon Musk's rocket company on Tuesday, marks one of the largest tech acquisitions in history and signals SpaceX's deepening push into enterprise artificial intelligence. The deal follows an unusual arrangement struck in April, when SpaceX secured the right to acquire Cursor for $60 billion or pay a $10 billion breakup fee, according to The Verge. The acquisition is expected to close in the third quarter, with Cursor becoming a wholly owned subsidiary of SpaceX, NBC News reported. Business Insider noted that the move reflects SpaceX's growing interest in integrating advanced AI capabilities into its operations, potentially for software development and engineering applications related to its space exploration and Starlink projects. "The acquisition follows a previous partnership in April where SpaceX collaborated with Cursor to develop AI models optimized for coding tasks, with SpaceX initially stating it would pay $10 billion for the collaboration." SpaceX's IPO, which took place just days before the acquisition announcement, valued the company at over $2 trillion and raised $85.7 billion, making it the largest listing in history, according to BBC. The IPO also made SpaceX the world's fifth most valuable company by Tuesday, surpassing Amazon, BBC reported. Cursor is a popular AI coding assistant tool that uses artificial intelligence to automate code writing, according to multiple outlets. The acquisition is seen as a move to strengthen SpaceX's enterprise AI capabilities and compete with rivals like Anthropic and OpenAI, The Verge reported. The deal is expected to close soon, according to an SEC filing cited by The Verge. "The IPO valued SpaceX at over $2 trillion and raised $85.7 billion in the largest listing ever." With this acquisition, SpaceX is positioning itself at the intersection of space technology and cutting-edge AI development, a strategy that could accelerate software innovation across its various projects, including Starlink and future Mars missions. The $60 billion price tag underscores the high value placed on AI coding tools in the current tech landscape.

7 sources · 20 Jun

Microsoft confirms Windows Recycle Bin bug after June security update, fix in progress

Microsoft has acknowledged a bug introduced by the June 9, 2026 security update (KB5094126) that causes the Recycle Bin to display internal system filenames instead of original file names when users permanently delete a single item. The issue affects Windows 11, Windows 10, and Windows Server, according to pureinfotech.com. "When users permanently delete a single item from the Recycle Bin, the confirmation dialog incorrectly displays an internal system filename (e.g., $Rxxxxx.ext) instead of the original filename." Microsoft has labeled the problem as cosmetic on its Windows Release Health Dashboard and confirmed that a fix is in development, pureinfotech.com reported. The bug was introduced by the KB5094126 update, which was part of the June Patch Tuesday rollout. Hacker News reported that the June 2026 Patch Tuesday update (KB5095051) for Windows 11 specifically has also introduced this Recycle Bin bug. Beyond the filename display issue, users have reported additional stability problems and OneDrive-related glitches since the update rolled out, according to Hacker News. "While Microsoft has officially acknowledged this specific issue, users have reported additional stability problems and OneDrive-related glitches since the update rolled out." The coverage from Hacker News highlights ongoing quality control concerns with Windows updates, noting that this is not the first time a Patch Tuesday update has introduced unexpected side effects. Microsoft has not yet provided a timeline for the fix, but the company has stated it is working on a resolution.

2 sources · 19 Jun

Snap's SPECS Glasses Arrive After a Decade of AR Development, Blending AI and Mixed Reality

Snap Inc. unveiled its long-anticipated augmented reality glasses, called SPECS, at the Augmented World Expo 2026 in Long Beach, California, marking the culmination of over a decade of work in the field. The lightweight device is designed to merge mixed-reality display technology with AI assistant features, aiming to make AR a practical part of daily life rather than a niche gadget. According to Hacker News, the glasses are intended to help users "create, connect, learn, and accomplish tasks while staying present in the real world," continuing Snap's vision of using technology to deepen human connection. Dezeen reported that the product represents Snap's effort to blend the immersive qualities of mixed-reality headsets with the practical, everyday form factor of AI glasses. "The product represents Snap's effort to blend the immersive qualities of mixed-reality headsets with the practical, everyday form factor of AI glasses." This positioning sets SPECS apart from bulkier competitors, as Snap aims to make AR a seamless accessory rather than a cumbersome headset. The glasses allow users to enjoy entertainment, work on the go, and access AI assistance through an integrated display, according to Dezeen. Hacker News noted that SPECS are designed to blend AI assistance, work tools, entertainment, and shared experiences into everyday life. The announcement at the Augmented World Expo signals Snap's confidence that the technology is finally ready for a broader audience, after years of iterative development and earlier prototypes like the Spectacles camera glasses. "The glasses are designed to help users create, connect, learn, and accomplish tasks while staying present in the real world." With SPECS, Snap is betting that consumers are ready to wear AR daily, a bet that will test whether the company can translate its social media dominance into hardware success. The product's lightweight design and focus on practical use cases could help it avoid the pitfalls that have plagued other AR efforts, but competition from Meta and Apple remains fierce in a market still searching for a breakout hit.

2 sources · 19 Jun

Ubisoft Co-Founder Claude Guillemot Dies in Western France Plane Crash

Claude Guillemot, a co-founder of the video game giant Ubisoft, died on Friday evening when the small plane he was piloting crashed in Western France. He was 69 years old. Ubisoft confirmed the news in a brief statement expressing deep sadness over the loss, as reported by Deadline. The accident occurred near the aerodrome of La Baule, a beach resort where Guillemot owned a vacation home. According to Deadline, the Cessna 421 twin-engine plane he was flying also killed his flight instructor in the crash. The exact cause of the crash has not yet been disclosed. Guillemot co-founded Ubisoft with his brothers in 1986, building the company into a global powerhouse known for blockbuster franchises such as Assassin's Creed. He also served as chairman of Guillemot Corp., the family holding company, as reported by Hacker News. "Ubisoft released a statement expressing deep sadness over the loss." The company's statement offered condolences to Guillemot's family, though it did not provide further details about the accident. The loss marks a somber moment for the French gaming industry, which Guillemot helped shape over nearly four decades.

2 sources · 20 Jun

Retro MMORPG EverQuest Legends Sets July 28 Launch With Subscription Model

Daybreak Game Company and developer Game Jawn have announced that EverQuest Legends, a reimagined version of the classic MMORPG, will launch for PC on July 28 at a base price of $19.99. According to Gematsu, the purchase includes a one-month subscription, with recurring monthly subscriptions costing $9.99. "a reimagined version of the classic EverQuest" Pre-orders are now available, offering access to a pre-order beta running from July 1 to July 21, character name reservation, and an exclusive in-game title, as reported by Gematsu. Bleeding Cool described the game as a "retro remake" aimed at recreating the original EverQuest experience for nostalgic players. The subscription model mirrors the original game's structure, with Gematsu noting that the base purchase includes the first month of access. Bleeding Cool emphasized that the title is designed to capture the feel of the classic MMORPG for longtime fans. "a retro remake of the classic MMORPG" With pre-orders now live and a beta period set for early July, players eager to revisit Norrath can secure early access and exclusive perks ahead of the full launch later this summer.

2 sources · 19 Jun

Garfield's dream-world platformer sets September 2025 launch despite busy release window

Garfield: Escape From Monday, a new platformer from developer Balio Studio and publisher Alias Systems, is officially launching on September 26, 2025, according to a new trailer that shows off the game's hand-drawn animation and voice acting. The game follows the lasagna-loving cat as he navigates a dream world to escape an endless Monday, with Jon Arbuckle and Odie joining him, Polygon reported. "The game follows Garfield as he navigates a dream world to escape a never-ending Monday." The trailer highlights gameplay mechanics including parkour, combat, and puzzle-solving, all wrapped in a hand-drawn art style. Balio Studio and Alias Systems are sticking with the September release date rather than delaying to avoid competition from other major games, Polygon noted. Meanwhile, a separate Garfield project was announced by Microids and Paramount: Garfield, Escape From Monday, a 3D platforming adventure set for September 2026 on PC, PlayStation, Xbox, and Nintendo consoles, Bleeding Cool reported. That game is triggered by Jon's spinach lasagna and features costume powers like turkey, surfer, and cowboy, with players exploring Garfield's biggest nightmares alongside Jon, Odie, and Arlene. "The game is set inside a nightmare sparked by Jon's spinach lasagna." Both titles share a similar premise of Garfield trapped in a nightmare world, but the 2025 release from Balio Studio and Alias Systems appears to be a 2D platformer with hand-drawn visuals, while the 2026 Microids version is a 3D adventure with costume-based abilities. Fans of the iconic cat will have two distinct ways to help him escape Monday, though they will need to wait until September 2025 for the first one.

2 sources · 19 Jun

Jessie Takes the Lead as Toy Story 5 Eyes Record-Breaking $275M Global Debut

Disney and Pixar's 'Toy Story 5' is off to a roaring start at the global box office, with projections pointing to a franchise-record opening weekend. The film, which focuses on the beloved character Jessie, earned $71 million on its opening Friday from 4,425 North American theaters, according to Variety. That strong first day sets the stage for a domestic weekend haul between $150 million and $175 million, which would easily surpass the previous franchise high of $120 million set by 'Toy Story 4'. "Toy Story 5 is projected to open with a franchise-best $275M worldwide, including $140M domestically at 4,425 theaters." Deadline reported that the global projection of $275 million includes a domestic estimate of $140 million, making it the biggest domestic start of 2026 so far. The film's success is part of a broader summer box office resurgence, with North American ticket sales reaching nearly $1.6 billion this season, according to Deadline. The strong performance underscores the enduring appeal of the 'Toy Story' franchise, which first launched in 1995. Variety noted that the film's focus on Jessie's story marks a shift from the previous installments' emphasis on Woody and Buzz Lightyear. This narrative choice appears to be resonating with audiences, as evidenced by the record-breaking projections. The opening weekend figures, if realized, would cement 'Toy Story 5' as the most successful launch in the series' history. With a robust start and positive word-of-mouth expected, 'Toy Story 5' is poised to continue Pixar's streak of box office dominance. The film's performance also bodes well for the remainder of the summer movie season, which has already seen significant momentum.

2 sources · 20 Jun

Amazon MGM Drops Nearly Finished Sam Altman Biopic After OpenAI Partnership

Amazon MGM Studios has dropped Luca Guadagnino's nearly completed film "Artificial," a biopic about OpenAI CEO Sam Altman starring Andrew Garfield, just months after Amazon announced a major partnership with the AI company. The decision, made by Prime Video and Amazon MGM Studios chief Mike Hopkins, was communicated to the filmmaker and producing team, according to The Hollywood Reporter. The film, which was originally slated for an early 2027 release according to Hacker News, will now be shopped to other studios. "The film will be better served if it were released by a different studio." This statement from Amazon, as reported by Rolling Stone, underscores the conflict of interest created by the company's deepening ties with OpenAI. The Verge noted that the film covers the dramatic five-day period in 2023 when Altman was fired and reinstated as CEO, a story that now clashes with Amazon's corporate strategy. Amazon's partnership with OpenAI, announced in February, includes a $50 billion investment and expanded use of Amazon Web Services, as detailed by Variety. Engadget reported that the $38 billion cloud deal and the investment created a clear conflict, leading to the film's removal from the studio's release schedule. Gizmodo highlighted that the nearly finished movie is now in limbo, raising questions about how corporate investments can influence creative projects. The film, directed by Luca Guadagnino, also features Monica Barbaro, Ike Barinholtz, and Yura Borisov in key roles, according to The Verge. As the project seeks a new home, the situation illustrates the tension between Amazon's business relationship with OpenAI and its prior commitment to the film, a dynamic that Rolling Stone explored in depth. The future of "Artificial" remains uncertain, but its journey reflects the growing intersection of big tech investments and Hollywood storytelling.

7 sources · 20 Jun

Famke Janssen Playfully Calls Out Marvel for Leaving Her Out of 'Avengers: Doomsday'

Famke Janssen, known for portraying Jean Grey/Phoenix in the original 'X-Men' trilogy, has publicly stated that Marvel Studios made a misstep by not casting her in the upcoming 'Avengers: Doomsday.' The actress made the remarks at Spacecon 2026, according to Variety, adding a lighthearted tone to her critique. "Marvel made a mistake." Janssen's comments come as several of her former 'X-Men' co-stars are reportedly returning for the film, according to Deadline. While she did not specify which actors are involved, she expressed her belief that she should have been brought back as well, though she acknowledged a personal flaw that might complicate such a role. "I'm bad at keeping secrets. My face gives everything away." In her remarks at the convention, Janssen humorously admitted she struggles with confidentiality, noting that her facial expressions betray her, as reported by Variety. She also joked about her acting abilities, framing her criticism as playful rather than bitter. Despite her exclusion, Janssen's comments reflect a broader fan interest in seeing legacy characters from the Fox 'X-Men' universe integrated into the Marvel Cinematic Universe. 'Avengers: Doomsday' is expected to feature a multiverse-spanning storyline, making such crossovers plausible, though Marvel has not confirmed any casting decisions beyond the initial announcements.

2 sources · 20 Jun

UK government to restrict under-16s from social media, citing child safety concerns

The UK government is set to announce a sweeping ban on social media access for children under 16, Prime Minister Sir Keir Starmer confirmed on Monday. The measure, which covers major platforms including TikTok, Instagram, Facebook, X, YouTube, Snapchat, and Reddit, represents one of the most aggressive moves by a Western nation to regulate children's online activity, according to Channel News Asia. "world leading action" Starmer's proposal goes further than Australia's landmark ban enacted in December 2025, the BBC reported. In addition to barring under-16s from social media, the UK plan includes restrictions on chatbots, a curfew for older teenagers to prevent late-night scrolling, and blocking children from livestreaming on "safer" sites or talking to strangers on gaming apps, according to Hacker News and the BBC. The announcement follows a nationwide consultation and comes ahead of a G7 summit in France, where Starmer will face political challenges including his defense secretary's resignation, ByteS EU reported. The prime minister framed the move as a choice between supporting families and maintaining ineffective existing protections, NBC News noted. A government survey cited by Deadline showed 90% public support for the ban. Starmer acknowledged the implementation will not be easy but emphasized the UK government has agency in pushing back against the power of big technology companies, Channel News Asia reported. The children's tsar has called for the ban to extend to under-18s, Kent Online added. Further details are expected next month, the BBC reported, as the UK follows Australia's lead in introducing the world's first outright social media ban for under-16s in December 2025. The measures also include a separate consultation to produce guidance for parents on when children should get their first smartphone and recommended screen time limits for under-16s, according to Kent Online.

7 sources · 20 Jun

Anthropic Pulls Two AI Models Worldwide After U.S. Export Order Targets Foreign Nationals

Anthropic has disabled global access to its Fable 5 and Mythos 5 AI models after the U.S. government issued an emergency export control directive citing national security concerns. The order, which arrived at 5:21pm ET, suspends access for any foreign national, including foreign national employees of Anthropic, both inside and outside the United States. Anthropic complied by taking both models offline for all customers, while other models remain unaffected, according to anthropic.com. "The directive was received at 5:21pm ET and lacks specific details about the national security concerns behind it." The move marks the U.S. government's most significant step to date in restricting access to advanced AI models, as reported by the San Diego Union-Tribune. Anthropic had released a limited version of Fable 5 widely just this week, while the more advanced Mythos 5 had tightly restricted access due to cybersecurity concerns. The suspension came just three days after Fable 5's public launch, forcing the company to abruptly disable access for all users, according to Le Monde. According to Bloomberg, the directive was issued under the Trump administration's export control authorities. The order affects even Anthropic's own foreign national employees, a detail that undercodetesting.com described as unprecedented: the first time the U.S. has banned a company's own employees from accessing its technology based on nationality. The action was triggered by a claimed "jailbreak" that reportedly exposed minor vulnerabilities, according to ground.news. Anthropic has publicly stated it disagrees with the government's handling of the matter, as reported by the San Diego Union-Tribune. The company argues that the directive is based on a misunderstanding, stating the identified jailbreaking technique revealed only minor vulnerabilities, according to ground.news. While the practical impact is limited due to the models' recent release and limited availability, the move represents a major escalation in treating frontier AI systems as sensitive defense assets subject to real-time kill switches, as undercodetesting.com noted. "This represents a major escalation in treating frontier AI systems as sensitive defense assets subject to real-time kill switches, affecting even the company's own employees and foreign nationals worldwide." The long-term ripple effects of this policy remain uncertain, but the directive signals a new era of government oversight over advanced AI development and access.

7 sources · 20 Jun

Anthropic ordered to pull top AI models in sudden government crackdown, fueling global push for independent AI

The Trump administration ordered Anthropic to shut down its most advanced AI models, Fable 5 and Mythos 5, over the weekend, blocking access for all foreign nationals including the company's own employees. The sudden, sweeping action with little warning has sparked concerns about US dominance over frontier AI technology and has accelerated sovereign AI development efforts in other countries, according to The Verge. "The sudden, sweeping action with little warning has served as a wake-up call for other nations about US dominance over frontier AI technology." This incident highlights the growing tension between national security interests and the global nature of AI research, as countries now scramble to reduce dependence on American-controlled AI systems. Disagreements over Anthropic's Fable 5's guardrails and a workaround led to the company's most advanced systems being taken offline, NBC News reported. The government gave Anthropic only 90 minutes' notice before forcing the withdrawal of Mythos and its commercial version Fable from the market, according to The Washington Post, which framed the action as a dangerous omen driven by grudges rather than sound policy. "The government's approach to AI regulation is driven by grudges rather than sound policy." Anthropic employees and industry experts allege the government's actions go beyond reasonable oversight into targeted and unreasonable interference, Gizmodo reported. The move has sparked debate about government overreach in AI regulation, with many questioning the administration's motives and the capricious nature of the ban. The incident serves as a wake-up call for nations dependent on American AI infrastructure, according to buff.ly, as countries worldwide accelerate their sovereign AI development efforts to insulate themselves from similar disruptions in the future. The long-term implications for global AI competition and collaboration remain uncertain as the industry grapples with this unprecedented government intervention.

5 sources · 19 Jun

Voice of Disney's Lilo and 'The Ring' Star Daveigh Chase Dies at 35

Daveigh Chase, the actress who voiced the beloved title character in Disney's "Lilo & Stitch" and played the iconic Samara Morgan in the horror film "The Ring," has died at age 35. Her boyfriend Roy Hernandez confirmed her death to TMZ, according to multiple outlets, after she was hospitalized with meningitis and serious blood infections. "Her boyfriend Roy Hernandez announced her death to TMZ, having recently set up a GoFundMe page for her medical expenses after she was diagnosed with meningitis and serious blood infections." Chase's death resulted from complications of meningitis, as reported by Deadline, and she was also being treated for a blood infection. Variety added that she had been hospitalized for malnutrition in Los Angeles before her passing. Chase won an Annie Award for her voice acting as Lilo Pelekai in the 2002 animated film "Lilo & Stitch," according to Variety. The A.V. Club noted that her career spanned from 2001 to 2010 and included work on influential horror movies, Studio Ghibli translations, and Disney animation. "Chase won an Annie Award for her voice acting as Lilo Pelekai in the 2002 animated film." Her role as Samara Morgan in the 2002 horror film "The Ring" became one of her most recognized performances, with Decider highlighting her chilling portrayal of the character who crawls out of a television. TVLine reported that she was best known for voicing the title character Lilo in Disney's "Lilo & Stitch" franchise. The news of her passing at such a young age has been met with an outpouring of tributes from fans and colleagues, remembering her contributions to both animated classics and horror cinema.

5 sources · 20 Jun

J.K. Simmons Leads MGM+ Crime Drama 'The Westies' Set in 1980s Hell's Kitchen

MGM+ has released the first trailer and announced a July 12 premiere date for its upcoming crime drama 'The Westies,' starring J.K. Simmons. The series, co-created by 'Narcos' alum Chris Brancato, delves into the violent world of an Irish-American organized crime gang operating in early 1980s Hell's Kitchen, as reported by Deadline. According to Decider, the show also stars Titus Welliver and is positioned as a gritty period piece for fans of 'Godfather of Harlem,' 'Boardwalk Empire,' and 'Narcos.' The series explores power struggles, rivalries, and brutal violence within the infamous gang, which competes with the Five Families for control of construction contracts tied to the Jacob Javitz Convention Center, Deadline noted. "The Westies" is set in early 1980s Hell's Kitchen and follows an Irish-American organized crime gang competing with the Five Families for control of construction contracts tied to the Jacob Javitz Convention Center. This backdrop of turf wars over lucrative city projects adds a layer of historical specificity to the drama, grounding the gang's conflicts in real New York development battles of the era. The premiere date of July 12 gives viewers a summer target for the new series, which promises to deliver the intense, character-driven storytelling Brancato is known for from his work on 'Narcos.'

2 sources · 20 Jun

TV Comedy Legend James Burrows, Director of Over 1,000 Episodes, Dies at 85

James Burrows, the legendary television comedy director and co-creator of the iconic sitcom "Cheers," has died at age 85. According to multiple outlets, his family confirmed his passing, with TVLine reporting that he died in his sleep on a Friday. Deadline noted that his death followed a brief illness, while Rolling Stone stated he was surrounded by loved ones and that no cause of death was provided. Burrows directed and shaped some of the most beloved sitcoms in television history, including "Taxi," "Friends," "Frasier," "Will & Grace," "Two and a Half Men," and "The Big Bang Theory," according to Deadline. Rolling Stone reported that he helmed over 1,000 episodes across five decades, making him one of the most prolific directors in the medium. "the greatest of all time." That tribute, reported by Deadline from colleagues including Lisa Kudrow, Tony Danza, and Sean Hayes, reflects the widespread admiration Burrows earned over his career. IndieWire noted that his family announced his peaceful passing and celebrated his five-decade career as one of television's most influential and beloved directors. Variety reported that his family issued a statement to People celebrating his career as one of television's most influential comedy directors who brought joy to audiences worldwide. Rolling Stone remembered him as a mentor and creative force who shaped generations of comedy, while Deadline highlighted him as the most decorated director in multi-camera format television. Burrows' legacy as a co-creator of "Cheers" and a director who defined the multi-camera sitcom format will endure. As TVLine noted, he was one of the most influential directors in television comedy history, known for shaping some of the most beloved sitcoms of all time.

6 sources · 20 Jun

Vercel's New Open-Source Framework Eve Aims to Standardize AI Agent Development

Vercel has introduced eve, an open-source framework for building, running, and scaling AI agents, now available in public preview. The framework treats an agent as a simple directory of files, with the smallest agent requiring just two files: a model and instructions. Adding tools, skills, channels, or schedules is done by adding files, which eve automatically wires in at build time without boilerplate, according to Vercel. "eve draws a parallel to how Next.js standardized web development, positioning eve as a similar standardizing force for agent development." This comparison underscores Vercel's ambition to make eve as foundational for agents as its own Next.js became for web apps. The framework includes built-in features like durable execution, sandboxed compute, approvals, and evals, aiming to eliminate the need for developers to hand-roll common infrastructure, Vercel reported. Users can scaffold and start a new agent with a single command that installs dependencies, scaffolds the project, and starts a dev server, Vercel added. This streamlined setup could lower the barrier for developers experimenting with AI agents in production environments. By bundling essential capabilities such as tracing, channels, and approvals directly into the framework, eve seeks to reduce the complexity of deploying agents at scale. Vercel's announcement positions the tool as a comprehensive solution for teams looking to move from prototype to production without building custom infrastructure from scratch.

2 sources · 20 Jun

Chinese AI Lab Z.ai Releases Open-Source GLM-5.2 With Million-Token Context Window

Chinese AI lab Z.ai has released GLM-5.2, a 753-billion parameter Mixture-of-Experts model with a 1-million-token context window, now available under an MIT open-source license. The model, which features 40 active parameters and is text-only input, was first made available to coding plan subscribers on June 13th before a public release on June 16th, according to Simon Willison's blog. "Z.ai launches GLM-5.2 with 1-million-token context window; MIT-licensed release coming next week" Hacker News reported that initial access was limited to subscribers of the GLM Coding Plan, while the broader MIT-licensed release is expected next week. This marks a significant leap from the previous GLM-5.1 model, which had a 200,000-token context window. The release is positioned as a direct response to US export restrictions targeting Anthropic models, according to Pandaily. The outlet noted that the move showcases "China's push for AI self-reliance and open-source alternatives amid escalating tech tensions between the US and China." "The key challenge addressed is making long context engineering-usable by ensuring reliability under real engineering pressure, rather than simply accepting more tokens." Hugging Face highlighted that GLM-5.2 is designed specifically for long-horizon tasks, delivering substantial improvements over GLM-5.1 by maintaining quality across long, complex coding-agent trajectories. The model's massive context window could enable processing of extremely long documents or codebases in a single pass, as noted by Hacker News. Simon Willison's blog reported strong community buzz around the release, underscoring the model's potential impact on open-source AI development. With its MIT license and million-token capability, GLM-5.2 represents a notable advancement in making long-context AI models accessible to developers worldwide.

4 sources · 19 Jun

Open-weights AI model GLM-5.2 tops Intelligence Index with score of 51

Z AI's GLM-5.2 has claimed the top spot among open-weights models on the Artificial Analysis Intelligence Index, scoring 51 and surpassing rivals like MiniMax-M3 and DeepSeek V4 Pro, according to Hacker News. The model, which shares the same architecture as GLM-5.1 with 744 billion total parameters and 40 billion active parameters, improved by 11 points over its predecessor. "GLM-5.2 becomes top open weights AI model on Intelligence Index with score of 51" This leap in performance underscores Z AI's focus on advancing scientific reasoning and other evaluation benchmarks, as reported by eicker.news. The model is available through Z AI's first-party API and third-party providers, maintaining similar pricing to GLM-5.1 at $1.4 per 1 million input tokens, $4.4 per 1 million output tokens, and $0.26 per 1 million cache hit tokens, according to Hacker News. "The model shows notable improvements in scientific reasoning and other evaluations" With a score of 51, GLM-5.2 outperforms competitors MiniMax-M3 and DeepSeek V4 Pro, both of which scored 44, as noted by both outlets. This positions Z AI as a strong contender in the open-weights AI space, offering a powerful model at competitive pricing.

2 sources · 19 Jun