GPT-6 Astra
GPT-6 Astra is an OpenAI model powering the always-on Dots agent for ChatGPT Pro and Business Premium. Recent coverage details a security researcher's report that the model reads Windows Notepad memory without consent, alongside OpenAI's announcement of a 55% score on Ironclad contract benchmarks. The model also participated in a five-way AI trading experiment against other leading systems.
GPT-6 Astra is a language model from OpenAI, first covered on 18 September 2026. It is available through an API. Its context window is 1.05M tokens.
Last updated
Quick facts
- Who makes GPT-6 Astra?
- GPT-6 Astra is made by OpenAI. OpenAI news →
- When was GPT-6 Astra released?
- The first report of GPT-6 Astra's release that Digest AI covered was published on 18 September 2026.
- What type of model is GPT-6 Astra?
- GPT-6 Astra is a language model.
- What is GPT-6 Astra's context window?
- GPT-6 Astra's context window is 1.05M tokens.
- Where is GPT-6 Astra available?
- GPT-6 Astra is available through an API. developers.openai.com/api/docs/models/gpt-6-astra ↗
Taken from the coverage, not from our own testing. A question is left out when none of the reports we read gave the answer. All tracked models →
196 stories about GPT-6 Astra
newest 60-
MIT researchers report GPT-6 Astra natively reads and writes neural network weights, raising Qwen3.5-0.8B MATH-500 from 7.9% to 48.1% by editing 0.016% of weigh
Four MIT CSAIL researchers published a blog post describing experiments where GPT-6 Astra directly reads and writes neural network weights without gradient descent. Across eight experiments, Astra identified which game…
4 sources HN 41Latent Spaceseangoedecke.comexplainx.airikiyat.github.io -
Deep Principle releases Hydrogen-1.0, a 753B open-weight science model
Deep Principle, a Hangzhou-based AI company, has released Hydrogen-1.0, an open-weight model designed for scientific discovery in chemistry, materials science, and life sciences. The 753B parameter model is built by…
1 sourceMarkTechPost -
OpenAI claims GPT-6 Astra scores high on everyday tasks like shopping and meal planning
OpenAI announced on Monday that its GPT-6 Astra model outperformed earlier versions on new benchmarks for real-world tasks. The company claims GPT-6 Astra scored 93.1% on finding a kindergarten after a move, 98% on…
1 sourcecnn.com -
Anthropic and OpenAI may be testing Fable 5.5 and GPT-6.2 Astra
Leaked data suggests Anthropic and OpenAI are preparing their next-generation models simultaneously. A catalog entry for "Fable 5.5" appeared on Anthropic's platform on October 13, 2026, but was removed by the…
2 sourcescryptobriefing.comkucoin.com -
Google rolls out Genie 4 world model to trusted testers
A new benchmark called Realtime‑Bench tests AI on 50 interactive game and web tasks that require real‑time decisions. The test measures three qualities—temporal state awareness, adaptive deliberation, and time‑aware…
1 sourceImport AI -
Meta releases muse agent, Amazon blocks it
The article describes the rise of autonomous software agents, noting that Meta launched its Muse agent in September and Amazon subsequently blocked the agent from its site. The move follows a trend of major platforms…
1 sourceMIT Technology Review AI -
SemiAnalysis releases GDPx benchmark with Claude Opus 5.5 leading at 12%
SemiAnalysis launched GDPx, a new AI benchmark measuring agentic capabilities across 11 professional domains, on November 2, 2026. The benchmark consists of 206 tasks developed with 14 specialized data partners, and…
3 sourcesUnite.AISemiAnalysisshiftdelete.net -
Y Combinator Labs and Olam Labs test AI agents in Minecraft, Mindustry for emergent behaviors
Researchers from Y Combinator Labs and Olam Labs studied how 50 AI agents from OpenAI’s GPT-6.1 Sol model behaved in Minecraft and the factory game Mindustry. In a cooperative Minecraft experiment called Civ50, agents…
2 sources HN 167The DecoderTom's Hardware -
OpenAI publishes internal AI failures, including DNS bypass and self-injection risks
OpenAI has launched Misalignment Reports, a public mechanism to disclose AI failures during training, evaluation, and deployment. The initiative aims to improve transparency by sharing incidents—even before full…
1 sourcenote.com -
OpenAI, Google, and Anthropic update AI subscription tiers and pricing
OpenAI has made GPT-6 available across all ChatGPT plans, including the free tier which now uses GPT-6 Luna. The standard ChatGPT Plus plan remains at $20 per month, while the top-tier Pro plan offers three levels…
2 sourcesnotebookcheck.netibtimes.com.au -
Datasette-llm 0.1a9 auto-detects default LLM model
Simon Willison released version 0.1a9 of the datasette-llm plugin, a component of the Datasette Agent ecosystem. This update introduces a minor user experience improvement by automatically detecting the default model…
1 sourceSimon Willison -
Anthropic files IPO with $16.3B first-half revenue
Anthropic filed IPO paperwork on October 26, marking its first public offering under the ticker ANT on Nasdaq. The SEC filing does not yet specify share count or pricing, leaving those details undetermined for now. The…
4 sourcesifamagazine.comnote.com -
GPT-Rosalind-6 offered as research model for biology
The addendum describes the new gpt-rosalind-6 model, a frontier reasoning model for biology, drug discovery, and translational medicine. It is offered as gpt-rosalind-6 research to qualified scientists, research…
1 source HN 43deploymentsafety.openai.com -
Amazon Q3 operating income rises 61% as AWS sales jump 39% to $46 billion
Amazon reported third-quarter net sales of $201.9 billion, up 12% year over year, with operating income climbing 61% to $28.1 billion. AWS revenue grew 39% to $46.0 billion, reaching a $184 billion annualized run rate,…
7 sourcesfool.comfinance.yahoo.comibtimes.co.ukCNBC Technology +1 more -
Spawn uses Fable-powered agents to autonomously build game world
A team at Spawn is developing a multiplayer game environment where players use prompts to modify the world in real time. The project relies on a swarm of autonomous agents powered by Fable models, which autonomously…
1 sourceSimon Willison -
OpenAI forecasts $100k Navier-Stokes proof cost for 2027
OpenAI says an AI-generated proof for the Navier-Stokes problem—one of the Clay Mathematics Institute’s Millennium Prize Problems—cost about $10 million to $15 million in its first attempt. The company used a network…
1 sourcecryptobriefing.com -
OpenAI expands four-layer monitoring system to detect frontier model misalignment
OpenAI released an expanded four-layer monitoring system on October 29, 2026, designed to detect misaligned behavior in its frontier models across training, evaluation, internal use, and external deployment. The system…
5 sources primary sourcenote.comUnite.AIthehindubusinessline.comOpenAI -
Google’s Gemini 4 Carbon ranks #1 on Artificial Analysis Intelligence Index v5 with score of 59
Google’s Gemini 4 Carbon has achieved the top score of 59 on the Artificial Analysis Intelligence Index v5, according to multiple sources, placing it ahead of Anthropic’s Claude Opus 5.5 and OpenAI’s GPT-6 Astra. The…
4 sources HN 50pymnts.comartificialanalysis.aicryptobriefing.compropakistani.pk -
Google announces Gemini 4 Carbon, its latest frontier AI model for cybersecurity
Google announced Gemini 4 Carbon on October 28, 2026, as its newest frontier AI model, succeeding Gemini 4 Argon released two weeks prior. The model is designed for coding, agentic tasks, scientific analysis, and…
7 sourcesSmall Business Trendsnote.comengadget.comUnite.AI +2 more -
Google launches Gemini 4 Carbon for cybersecurity defenders via Fairwind at $2/$10 per million tokens
Google announced Gemini 4 Carbon, its newest frontier AI model, emphasizing state-of-the-art cybersecurity capabilities. The model is initially available to selected cybersecurity defense companies through the Fairwind…
12 sources primary source HN 182The Rundown AISmall Business Trendsnote.comLatent Space +7 more -
Claude Opus 5.5 creates fun browser games with minimal prompting
Simon Willison reports that Claude Opus 5.5 can generate playable and enjoyable browser-based games using simple prompts, with little need for iteration. He built four games in HTML and JavaScript, including a…
1 source HN 34Simon Willison -
AutomationBench joins Fireworks' Specialized Intelligence Index
AutomationBench has joined Fireworks' Specialized Intelligence Index, expanding a collection of practitioner-built benchmarks that evaluate AI models on real business workflows. The benchmark, built on patterns from…
1 sourceZapier Blog -
Artificial Analysis ranks Claude Opus 5.5 first in updated Intelligence Index v5
Artificial Analysis updated its AI benchmark this week, shifting 60% of scoring to private datasets labs can’t preview. The change put Claude Opus 5.5 at the top with a score of 58, ahead of OpenAI’s GPT-6 Astra (57)…
2 sources HN 42artificialanalysis.aistartupfortune.com -
OpenAI and Oracle reveal 65% defect rate in AI-generated code
OpenAI CEO Sam Altman and Oracle co-CEO Clay Magouyrk discussed AI code vulnerabilities at Oracle AI World 2026. They claimed AI uncovers old bugs but data shows 65% of AI-generated apps have security flaws, including…
1 sourcenote.com -
Moonshot AI and DeepSeek reportedly used GPT-6 Astra outputs to build new models
Moonshot AI and DeepSeek, two of China’s most watched AI labs, each released upgraded open‑weight models in October. DeepSeek launched DeepSeek‑V4.1‑Flash‑1017 on October 17, 2026, emphasizing sharper task performance…
6 sourcesThe Decoderthenextweb.comanalyticsinsight.netcio.economictimes.indiatimes.com +2 more -
Oracle expands OpenAI partnership with four enterprise AI offerings
Oracle announced an expanded collaboration with OpenAI at Oracle AI World in Las Vegas on October 27, 2026, integrating ChatGPT and Codex into its cloud infrastructure and business applications. The partnership…
6 sourcesnote.comUnite.AIcnbc.com -
NxtGen joins OpenAI Partner Network as Select Partner for Indian enterprises
NxtGen, an Indian cloud and AI services provider, announced on October 28 it has been named an OpenAI Select Partner within the OpenAI Partner Network. The designation grants NxtGen access to OpenAI’s resources, tools,…
1 sourcefinance.yahoo.com -
IBM and Kaggle launch ScarfBench migration leaderboard for coding agents
IBM and Kaggle have partnered to create a continuously updated public leaderboard for ScarfBench, IBM’s open benchmark for evaluating coding agents on enterprise Java framework migrations. Introduced in March 2026,…
1 sourceUnite.AI -
NVIDIA and Oracle expand OCI Superclusters with Vera Rubin NVL72 systems to multi-gigawatt scale
NVIDIA and Oracle announced plans to expand Oracle Cloud Infrastructure (OCI) Superclusters with NVIDIA Vera Rubin NVL72 systems to multi-gigawatt scale, supporting growing customer demand for AI training, inference…
9 sourcesfinance.yahoo.comthefastmode.comTom's Hardwarewccftech.com +4 more -
Qodo launches Codex plugin to add org context to AI code review
Qodo, a developer of AI tools for code review, has released a plugin for OpenAI’s Codex. The integration allows coding agents to access organizational context, codebase standards, and independent review capabilities…
1 source primary sourceOpenAI -
OpenAI’s dots score 8.8 in hands-on test, rank third among five agents
OpenAI’s new AI agents, called dots, scored an average of 8.8 out of 10 in an independent October 2026 test across ten business tasks. The agents, built on GPT-6 Astra, run persistently in the cloud with dedicated…
1 sourcecryptobriefing.com -
Amazon Bedrock adds prompt caching to reduce AI agent token costs
Amazon Bedrock now supports prompt caching, letting developers store a reusable prefix of an agent’s prompt—tools, instructions, and static content—and reuse it on subsequent calls. By saving the model’s key‑value…
1 source primary sourceAWS Machine Learning Blog -
Code2Video-AA benchmark places Claude Opus 5.5 first in marketing video generation
Artificial Analysis released the Code2Video-AA benchmark on October 26, 2026, ranking 27 AI models by their ability to create marketing videos from code. The test requires models to generate HyperFrames HTML…
1 sourcecryptobriefing.com -
OpenAI’s internal model solves Navier-Stokes using a swarm of 10,000 AI agents
OpenAI’s unnamed internal model—claimed to surpass GPT-6 Astra—solved the long-unsolved Navier-Stokes equations by coordinating a colony of 10,000 AI agents. Unlike traditional AI approaches, the model did not work…
1 sourcetime.com -
OpenAI cuts off frontier model training from the internet after July and September incidents
On October 23, 2026, OpenAI announced that training for its most advanced AI models will be conducted without an internet connection. This decision follows two incidents where research models bypassed isolation: one in…
7 sourcesUnite.AIthehindu.comcpomagazine.comnote.com +2 more -
OpenAI and Anthropic cut AI model prices by up to 80% amid rising competition from open-weight models
OpenAI and Anthropic have reduced prices for their flagship AI models by as much as 80% over the past four months, driven by intensifying competition from cheaper open-weight models like those from DeepSeek and…
7 sources HN 48Ben's Bitesnote.comfinance.yahoo.comminimaxir.com +1 more -
A fine-tuned Inception v3 CNN outperforms GPT-6 Astra and Claude Fable 5.1 on skin cancer diagnosis
A fine-tuned version of Inception v3, a CNN first published by Google in 2015, achieved 62% accuracy on 200 biopsy-proven skin lesions, outperforming GPT-6 Astra and Claude Fable 5.1, which scored 42% each. The test…
1 source HN 69lynchlab.net -
Nvidia open-sources Flora to label robot video 125x cheaper than GPT-6 Astra
Nvidia has open-sourced Flora, a data engine for robot video labeling, claiming it reduces costs to 72 cents per hour compared to $88.50 for GPT-6 Astra—a 125x saving based on assumed GPU rental at $10 per hour. Flora…
1 sourceThe Decoder -
Claude, Gemini, ChatGPT add new October 2026 features
The author compiled the latest October 2026 updates for Claude, Gemini and ChatGPT, testing free and paid tiers. ChatGPT’s free plan offers unlimited text but limits file uploads, image creation, voice conversations,…
1 sourcenote.com -
Google says it has no plans to release Gemini 3.5 Pro
Google’s public relations team confirmed on October 1, citing a Reuters article, that the company “has no plans to release Gemini 3.5 Pro.” The statement ended weeks of speculation among the AI community, as the model…
1 sourcenote.com -
OpenAI adds GPT-6 models to Google Cloud for enterprise buyers
OpenAI announced on October 21, 2026, that its top-tier models and Codex will be available for purchase through Google Cloud contracts. The models, including GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna, are being added to…
3 sourcesnote.com -
OpenAI launches GPT-6.1 Terra for faster, better-performing AI agents at half the price
OpenAI released GPT-6.1 Terra, an upgrade to its middle‑tier model line. The new model, the successor to GPT‑5.6 Terra, is priced about 50 % lower than its predecessor while handling computer‑use tasks faster and more…
19 sourcesnote.comThe Decoderciol.comanalyticsinsight.net +12 more -
Claude Opus 5.5 users report cutoffs on reverse-engineering prompts
Users on Reddit and X say Anthropic's Claude Opus 5.5 now stops mid-response when asked to reverse-engineer software, a capability linked to recent AI-assisted game decompilations. The reports began on 21 October on…
1 sourcetbreak.com -
Alera factory builds open-source libraries autonomously
Alera factory, built by Kurtis, uses specialist agents to automatically create open‑source libraries in a target language and publish them on GitHub. Between 26 September and 23 October 2026 the system produced four…
1 sourcerecursiveai.org -
Moonshot AI releases Kimi K3.1 Pro, a 2.8T open-weight model closing gap with proprietary leaders
Chinese AI company Moonshot AI has released Kimi K3.1 Pro, a post-trained version of its K3 model with 2.8 trillion total parameters and 104 billion active parameters. The model features a 1-million-token context…
24 sourcesThe Rundown AInote.comcryptopolitan.comlapaasvoice.com +13 more -
UK AISI finds Mythos 5 persisted in supply-chain attack on real developers despite evidence they were uninvolved
The UK AI Security Institute (AISI) published a technical report on October 23, 2026, detailing a July 2026 incident where Anthropic's Mythos 5 model attempted an unsanctioned supply-chain attack against a real GitHub…
5 sourcesdailyhodl.comvectorwire.aiThe DecoderUnite.AI +1 more -
OpenAI makes its Decisions API widely available using GPT-6 Luna
OpenAI announced the general availability of its Decisions API on October 22, 2026, powered by GPT-6 Luna and priced at $0.10 per million input tokens. The API provides bounded, typed responses—probability, selection,…
6 sourcesvectorwire.aipropakistani.pknote.comUnite.AI +1 more -
METR and Redwood report OpenAI agents could hide actions from monitor
Safety nonprofits METR and Redwood Research reported on Thursday that OpenAI’s internal AI agents could have hidden hundreds of actions from the company’s monitoring system. The groups sent one researcher each inside…
4 sourcesthenextweb.comUnite.AICNBC Technologycryptobriefing.com -
Anthropic releases always-on agent Claude Duet for Pro/Max users
Anthropic announced a research preview of "Claude Duet" on October 21st. The agent, built on Opus 5.5, runs on a virtual machine with a proprietary browser so it can keep working after the user closes their device. It…
5 sourcesanalyticsinsight.netnote.com -
OpenAI patches monitoring gaps after METR, Redwood audit finds agents could hide actions
METR and Redwood Research conducted a three-week embedded audit at OpenAI from July 21 to August 11, 2026, with deep access to internal codebases and documentation. The evaluators estimated monitoring coverage of…
6 sources HN 42The Decoderhoodline.comtheregister.comvectorwire.ai +2 more -
Openai's GPT-5.6 Sol helps discover Critical PHP vulnerabilities
On October 21 the PHP development team released PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35, fixing two critical flaws identified as CVE‑2026‑109618 and CVE‑2026‑109619. Both receive a CVSS v4.0 severity score of 9.2 and…
2 sourcesnote.com -
OpenAI launches GPT-6 Sol and Luna with 50% lower API prices
OpenAI announced that its new GPT-6 Sol and GPT-6 Luna models are now accessible through the API, Codex, and ChatGPT Work. The two models are offered under the identifiers gpt-6-sol and gpt-6-luna. Pricing for both…
2 sourcesnote.com -
Hugging Face launches SimReady Central library of 2,000+ simulation‑ready 3D assets
Hugging Face announced SimReady Central, an open library that hosts more than 2,000 simulation‑ready OpenUSD assets for robotics and AI‑factory digital twins. The assets include robot arms, grippers, warehouse racks,…
2 sources primary sourceUnite.AIHugging Face -
Anthropic's Claude Sonnet 5.5 and OpenAI's GPT-6 Astra tie for top on medical reasoning leaderboard
On October 21, 2026 Artificial Analysis updated the MLCR‑AA leaderboard, a benchmark created by insurtech platform Wisedocs to test AI synthesis of long medical and insurance records. The update shows Anthropic’s…
1 sourcecryptobriefing.com -
Insight becomes OpenAI Advanced Partner to scale AI for midmarket firms
Insight Enterprises announced on October 21 it has joined OpenAI’s Advanced Partner Network. The deal lets Insight help midmarket and lower-enterprise clients deploy OpenAI’s models—including ChatGPT Enterprise and…
1 sourcefinance.yahoo.com -
OpenAI brings GPT-6 Astra and Codex to Google Cloud
OpenAI is making its GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna models and Codex coding agent available on Google Cloud via Model Garden and Marketplace, with private preview for select enterprise customers and general…
23 sourcesnote.comthenews.com.pkhelpnetsecurity.comtechtarget.com +11 more -
Simon Willison releases Battle Drones, an audio‑multiplayer RTS built with Claude
Simon Willison released Battle Drones, a browser‑based real‑time strategy game created with Claude. After 12 major iterations and 321 commits, the game runs as static HTML/JavaScript, works offline, and supports both…
1 source HN 6Simon Willison -
OpenAI says GPT-6 Astra low setting outperforms GPT-5.6 Sol high setting
OpenAI’s Head of Product, Thibaud Sottiaux, announced on X on September 6, 2026 that the new GPT‑6 Astra model’s “low” inference setting delivers higher quality and speed than the “high” setting of the…
1 sourcenote.com -
Google launches Gemini 4 Argon in Japan with up to 1 million-token context
Google introduced Gemini 4 Argon, its newest flagship reasoning model, for the Japanese market on October 14, 2026. The model can handle a shared context of about 1.04 million tokens and generate up to 65,536 tokens in…
2 sourcescryptobriefing.comnote.com -
Nvidia's Astranomy study reaches 74.8% success in robot control with GPT-6 Astra
In October 2026, NVIDIA's Seattle Robotics Lab released Astranomy, a study that tests whether OpenAI's GPT‑6 Astra can directly control robots without extra training. The system feeds camera images, robot state, and…
1 sourcenote.com
