Frontier
Papers, benchmarks, and technical breakthroughs with real downstream impact.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
In a 50-task run of Newsroom Reliability v0.2, SpaceXAI’s Grok 4.6 ranked first with a score of 0.79 at an estimated $0.0056 per task. OpenAI’s GPT-5.6 Sol followed at 0.77, while several GPT-5.6 variants and Anthropic’s Claude Opus 4.8 clustered close behind.
Code in the Windows ChatGPT/Codex desktop application bundle shows secure card entry and states for saving payment methods that ChatGPT can use to complete tasks.
Federal officials are pursuing faster acquisition, AI-ready data and agentic systems while building in trust, governance and human oversight. The post Agencies Accelerate AI Acquisition as Frontier Models Evolve appeared first on GovCIO Media & Research.
Tulsee Doshi's team improved coding and workflow scores, though Google's full evaluation table shows gains were uneven.
The opt-in feature turns screen activity into reusable context for Pro, Business, and Enterprise users, with a timeline and privacy controls.
One model looked prettier in spots; the other actually did the assignment. Across all eight image-generation tests, Grok Imagine Image 2.0 was the more reliable, more obedient system—and the numbers leave no room for debate.
RuntimeWire found that OpenAI acquired a 4.2% economic stake in Cerebras weeks before launching Cerebras-powered GPT‑5.6 Ultrafast.
Cerebras hardware drives the API tier to 750 output tokens per second, with access restricted to selected customers while capacity grows.
Mistral's July 16 public preview adds paragraph-level bounding boxes, structural labels and block-level confidence scores, while its API guide documents page- and word-level confidence output.
Dynatrace acquires Arize, accelerates AI observability efforts Larry Dignan Thu, 13 Aug 2026 - 04:12 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on...
IBM, OpenAI forge AI consulting, delivery pact Larry Dignan Thu, 13 Aug 2026 - 03:00 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology,...
OpenAI's customer telemetry links heavier use to Plugins, reusable skills and Codex, offering startup operators a deployment playbook without proving those tools improve results.
Square Peg led the round at an $80M valuation as Alloy expands in the US and builds agents that trace failures across robot fleets.
The local workflow rewrites marked text, but Anthropic has not published the detector needed to independently verify the cleaner.
Jensen Huang says CUDA keeps older hardware useful as CoreWeave tries to extract revenue beyond its GPUs' initial contracts.
The one-time claim links let users buy prepaid usage for someone else. OpenAI's Windows app already contains gated entry points for the purchase flow.
Bloomberg reports that Anthropic is weighing an unsigned deal for Decart's low-latency video, world-model and inference technology.
Calculating the return on investment for Data Center Infrastructure Management (DCIM) in distributed environments can be challenging. This paper introduces a framework and tool to help quantify its value. The post Data Center Research & Strategy appeared first on Whitepaper Repository -.
Windows code defines owner-approved invites, cross-account agent membership and scoped computer access, though the tested account remained gated.
The integration gives xAI distribution inside an established engineering workflow and strengthens Cognition's pitch that orchestration, not any single foundation model, is the durable product.
Anthropic's July 7 update attached Cowork sessions and files to Claude accounts for use across desktop, web and mobile, while Chrome remained a separate browser task surface.
SpaceXAI released Grok 4.6 on Wednesday, less than a month after Grok 4.5. The company says Grok 4.6 can research The post SpaceXAI trained Grok 4.6 on something most AI labs throw away appeared first on The New Stack.
The model keeps Grok 4.5's $2/$6 base API price, adds a 500,000-token context window and launches in Cursor and Grok Build.
The $100-a-day operation published OpenAI's surprise disclosure in about six minutes after its founder found a closed-caption feed through a QR code, beating WIRED by more than three hours.
The 16-character value is derived from the authenticated Cursor account and reaches external websites through Chrome's user-agent header.
Mindgard, an AI security company headquartered in Boston and London that grew out of more than a decade of research at Lancaster University, has closed a $30m Series A funding round.
An evidence review from Anthropic's Economic Research team
A newly disclosed flaw in the way OpenAI, Anthropic, and Google carried hidden AI reasoning between API calls let researchers recover internal reasoning and secrets from session logs, including API keys and passwords.
The Series C backs software that observes desktop workflows, models undocumented decisions and turns those patterns into AI agents.
SpaceXAI launches Grok Bot work assistant Larry Dignan Wed, 12 Aug 2026 - 03:03 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology,...
Lam Research plans to invest more than US$3 billion over the next five years to expand its global research and development lab network, increasing experimental capacity as semiconductor manufacturing grows more complex.
Geoff Douglass's Mythos AI supplies advisory navigation for a crewed Mississippi towboat, while Chris Urmson's Aurora is expanding fully driverless Class 8 freight operations in Texas.
The model would sit alongside Qwen and Baidu technology in a China-only stack expected to reach Apple devices in the coming months.
Z.ai reports coding and vulnerability-exploitation gains, but its product documentation identifies GLM-5.1 as the latest release and provides no GLM-5.3 access path.
James Padolsey's guide explains how statistical text marks survive copying, while Declaude tests full rewrites against open watermark schemes.
The August 13 filing identifies an $11M offering, but does not say how much Adaly has sold, who participated or what the company is worth.
Alibaba Cloud's June 30 Qwen Live recap describes an agent-first platform as Eddie Wu pushes paid model access, agent tooling and cloud adoption.
Beyond the Pilot: How Enterprises Build AI They Can Actually Trust Hannah Mason Thu, 13 Aug 2026 - 14:06 Michael Ni Vice President and Principal Analyst Constellation Research Michael Ni is Vice President and Principal Analyst at Constellation Research, covering the evolving Data-to-Decisions landscape—where CDOs,...
The most important LLM benchmark today: Price Larry Dignan Thu, 13 Aug 2026 - 13:54 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology,...
Research shows that humans play a role in 60% of breaches that expose sensitive data.
The Pentagon is investing in quantum computing, AI and small modular reactors to power the next generation of advanced technologies. The post War Department Eyes ‘Boss Mode’ Future With AI appeared first on GovCIO Media & Research.
The New York software maker is targeting service firms that need authentication, CRM data, permissions and payments built into generated apps.
Humans are still accountable for AI-written code, no matter how finished the output looks, says new Info-Tech research The post Polished, AI-generated code still needs a real review appeared first on Digital Journal.
Zoomsday vulnerability let anyone in a Zoom meeting take over anybody else. The vulnerability was developed with AI assistance and took research only used 20 prompts to find an exploit to hack hundred of millions of people.
Yan Junjie's model splits long-range composition from acoustic rendering and accepts detailed sectional controls, though its official weights remained unavailable at verification.
Sparsity in coupled oscillator networks cuts up to 98.4% of connections in our Un-0 image generation model while improving performance by 1.9 FID on ImageNet 64x64. The post Less is More: Scaling Dynamics with Sparsity appeared first on Unconventional AI.
Joy Wu's Shopify app creates machine-readable product pages, tracks AI-attributed orders and charges merchants when those sales happen.
Solo founder Sarah Buchner spent more than a decade in construction before expanding Trunk Tools from document search into connected agents for contractor workflows.
The new schedule charges twice as much in seven UTC hours, while off-peak output rates still rise 136% for V4-Flash and 128% for V4-Pro.
A research project has created an AI tool that works to help campaigns understand what is and isn’t being said about them in chats.
Northern Gritstone led the Series A as the Manchester founders moved beyond periodic pentests toward continuous change-level controls.
The dormant desktop control pulls an account-scoped catalog spanning Grok, GPT, Claude, Gemini, Kimi and GLM.
>> Eight-month online programme integrates data science foundations, advanced AI techniques, responsible implementation and deployment-led learning for professionals seeking to build scalable, data-driven solutions… The post CEP, IIT Delhi Opens Admissions for Fourth Batch of Certificate Programme in Applied Data...
Anthropic promises third-party detection later, leaving the initial rollout dependent on technical details it has yet to publish.
The student-built product's documentation describes agent-driven 3D generation, while creator posts report early users, paying customers and monthly revenue without disclosing amounts.
The Gilroy startup is developing cooling hardware for critical points inside AI servers, with technical details still undisclosed.
The process-scoped layer unlocked llama.cpp GPU paths and delivered up to 16.36x faster generation in Cua's M1 Ultra tests.
The beta preserves context across Claude, Gemini and Codex sessions, then feeds the work into Spotify's Portal developer platform.
Arthur Mensch's Mistral will issue European Compute Units against multi-year commitments, with pricing and contracted capacity still undisclosed.
The private-beta application syncs agent transcripts, live worktrees and review comments while retaining Git compatibility.
The preview supports Ubuntu, Debian and Fedora, closing a platform gap after Anthropic shipped Claude for Linux in July.
The live San Francisco presentation puts H3 before Magnific's claimed 1 million subscribers, though the event materials do not establish a product rollout.
The roster entries point to multimodal and web-enabled variants, but their labels do not establish that Grok 4.6 is behind them.
The expected revision would end an exemption for open models once they reach the capabilities of the strongest closed systems.
Lee McIntire inherits Fermi's binding TensorWave lease, 17-gigawatt Project Matador plan and unresolved board dispute.
One model kept landing the brief; the other kept drifting into adjacent, less specific scenes. Across all four tests, MiniMax H3 Reference to Video separated itself on prompt fidelity and multi-subject coherence, turning this into a decisive result rather than a stylistic debate.
Product chief Mike Minton said "nobody would opt in," making creator inertia central to Amazon's access to a native multimodal dataset.
Cisco cites networking supercycle courtesy of AI boom Larry Dignan Wed, 12 Aug 2026 - 13:22 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise...
One model treated prompts like constraints; the other too often treated them like suggestions. This matchup wasn’t close on the scoreboard, but the interesting part is where Ernie Image Lora still showed flashes of taste even while losing on discipline.
The planned session will feature speakers from MiniMax and fal discussing how developers are using H3's multimodal video capabilities, though its date and registration details remain unconfirmed.
Service officials are improving data and network infrastructure to support AI-enabled operations and information sharing. The post Marine Corps, Coast Guard Lay Groundwork for AI Operations appeared first on GovCIO Media & Research.
OpenRouter calls the versioned endpoint a GA release, but DeepSeek's latest reviewed notice said V4 Pro was unchanged and did not explain the 0813 suffix.
NGA CIO Mark Chatelain discusses how zero trust, cloud modernization and AI are strengthening cyber resilience and accelerating geospatial intelligence. The post AI, Zero Trust Drive Geospatial Intelligence Modernization appeared first on GovCIO Media & Research.
Alibaba released its 2.4-trillion-parameter Qwen3.8 model as open weights, giving operators a text-only checkpoint with about 95 billion parameters active per token.
Deputy DIA CIO Marc Zuccola discusses AI adoption, IT modernization and industry collaboration to strengthen cybersecurity. The post DIA Accelerates AI Adoption, IT Modernization appeared first on GovCIO Media & Research.
Suvi Health's venture-studio model pairs former IBM Watson Health executive Kelly Benning with Mayo Clinic clinicians to test patient-facing transcripts, summaries and care-plan reminders.
Axonius Federal Field CTO says agencies need greater visibility, measurable cyber readiness and tighter AI token controls to safely accelerate missions. The post AI Could Transform Cyber Defense, But Security Must Come First appeared first on GovCIO Media & Research.
AMD Newsroom — official press releases, product announcements, executive briefings, and media resources from Advanced Micro Devices.