# [AINews] OpenAI to reach AGI bar by end-2026 > Source: > Published: 2026-08-28 07:12:10+00:00 Normally we eschew AGI timeline talk on Latent Space, because it is so ill defined and unaccountable, but, well, **missing it** would probably be the worse sin at this point. We last checked in on [OpenAI AGI timelines 9 months ago](https://www.latent.space/p/agent-labs), and, right on target, Chief Scientist Jakub Pachocki is now saying the unreleased Astra model is the “**Automated AI Research Intern**” he had aimed for by September 2026. Sama goes further in [their TIME interview](https://time.com/article/2026/08/26/openai-sam-altman-interview/?utm_source=twitter&utm_medium=social&utm_campaign=editorial&utm_content=260826) and estimates they’ll declare AGI achieved internally by December 2026. Start the clock. AI News for 8/22/2026-8/24/2026. We checked 12 subreddits, [544 Twitters]and no further Discords.[AINews’ website]lets you search all past issues. As a reminder,[AINews is now a section of Latent Space]. You can[opt in/out]of email frequencies! **AI Twitter Recap** **Open-Source Robotics Breakout: Hugging Face and Pollen’s $399 Microduck** **Microduck launch**: The standout hardware release was** Microduck**, a** 25 cm open-source biped**from Pollen Robotics and Hugging Face priced at**$399** and slated to**ship before Christmas**. It can be** trained in simulation and deployed on the real robot**, with** 15 actuators**and a notably rich sensor stack including** camera, speaker, LiDAR, NFC, Bluetooth, and Wi‑Fi**. Launch posts from[@pollenrobotics](https://x.com/pollenrobotics/status/2092915032052879425),[@Thom_Wolf](https://x.com/Thom_Wolf/status/2092923071829049592), and[@ClementDelangue](https://x.com/ClementDelangue/status/2092931447644442635)emphasize reinforcement-learning-based customization plus several pre-trained policies out of the box.**Why it matters technically**: The interesting part isn’t just “cheap cute robot,” but the package design: an** open simulator**, transfer from sim to hardware, and a form factor cheap enough to invite community policy training rather than just demo consumption. The simulator is already public via a Hugging Face Space, highlighted by[@HuggingApps](https://x.com/HuggingApps/status/2092994724214743063), and this open-loop from community training to real deployment is what got multiple researchers immediately buying units, e.g.[@yacineMTB](https://x.com/yacineMTB/status/2092962380816744788)and[@gneubig](https://x.com/gneubig/status/2092971650803208247).**Early traction and community experimentation**: The release resonated unusually broadly for robotics. Thom Wolf shared experiments such as a quick image-detector integration to let the robot**follow a laser pointer** in real time[@Thom_Wolf](https://x.com/Thom_Wolf/status/2092959363992326236), then reported sales velocity of**one Microduck every 5 seconds** and later**$1M in sales**[@Thom_Wolf](https://x.com/Thom_Wolf/status/2093014172531339383),[@Thom_Wolf](https://x.com/Thom_Wolf/status/2093023975173431449). The combination of low price, open sim, and embodied RL makes this one of the more credible “consumer-scale physical AI” launches in recent memory. **GLM-5.3-Flash/Ox Alpha Reveal and Local Open-Model Momentum** **Ox Alpha unmasked as GLM-5.3-Flash**: One of the biggest model stories was the confirmation that the mystery model** Ox Alpha**was actually** Z.ai / Zhipu’s GLM-5.3-Flash**, as noted by[@theo](https://x.com/theo/status/2093078228491731177),[@UnslothAI](https://x.com/UnslothAI/status/2092986464196002094), and[@togethercompute](https://x.com/togethercompute/status/2093015257560281099). The disclosed spec repeatedly cited across tweets:**320B total params, 18B active**,** 1M context**, and** hybrid attention**, with strong results on coding/agentic benchmarks.** Open weights + quantization + local serving**: The release caught attention because people quickly pushed it into local workflows. Unsloth said the model can run** 3-bit GGUF on 128GB RAM**[@UnslothAI](https://x.com/UnslothAI/status/2092986464196002094), while[@danielhanchen](https://x.com/danielhanchen/status/2092996385302094189)claimed**4-bit retains 93% accuracy** and makes the model practical on a**256GB Mac** or**two DGX Sparks**. This is exactly the kind of post-release ecosystem response open-model engineers care about: quantization, serving recipes, and real deployment constraints moving almost immediately.**Price/performance narrative**: Several tweets framed GLM-5.3-Flash as a new efficiency frontier.[@togethercompute](https://x.com/togethercompute/status/2093015257560281099)said it nearly matches Luna on DeepSWE while doing**more than twice as much work for the same budget**;[@theo](https://x.com/theo/status/2093069233571942510)called it good enough to reorder his model rankings;[@zainhas](https://x.com/zainhas/status/2093125213361938621)suggested using**high** rather than**max** reasoning effort because accuracy stayed roughly flat while token usage doubled. Baseten also highlighted**122+ TPS** serving throughput on day 0[@baseten](https://x.com/baseten/status/2093086722196172825), while Databricks cited**270 tok/s** and**10% higher quality than GLM-5.2 at 1/10 the cost** on OfficeQA Pro v2[@Yuchenj_UW](https://x.com/Yuchenj_UW/status/2093177892356472978). **Video Generation Race: Gemini Omni 1.1 Flash and H3 Max** **Gemini Omni 1.1 Flash**: Google released** Gemini Omni 1.1 Flash**, a multimodal video generation/editing model with several developer-facing controls:** scene extension to 40s**,** first/last frame control**,** 3-second video references**,** 360p draft mode**, and** 4K upscaling**. The rollout was announced by[@Google](https://x.com/Google/status/2093008576487072064),[@GoogleAIStudio](https://x.com/GoogleAIStudio/status/2093008678118998298), and summarized with prompting guidance by[@_philschmid](https://x.com/_philschmid/status/2093012878211072183). The most notable product detail is that Google is exposing increasingly explicit temporal and reference conditioning rather than just “prompt harder.”**Early leaderboard results**:[@arena](https://x.com/arena/status/2093015572212846673)reported Omni 1.1 Flash landing**#1 in Text-to-Video Arena** and**#2 in Image-to-Video Arena**, with a**+20 pt** lead over the #3 text-to-video model and a**+25 pt** improvement over prior Gemini Omni Flash on image-to-video. That does not settle all qualitative questions, but it indicates Google’s latest post-training and control stack is translating into preference data.**fal + MiniMax H3 Max**: In parallel, fal launched** H3 Max**with MiniMax, advertising** 15s of high-quality video in 5s**and “** 50x faster**” generation than other high-quality models[@krea_ai](https://x.com/krea_ai/status/2092990757506322661), with technical writeups from[@fal](https://x.com/fal/status/2093068605114204456)and praise from[@MiniMax_AI](https://x.com/MiniMax_AI/status/2093092333378224185). The theme across both launches is clear: inference optimization and productized controllability are now as important as base-model quality in video. **Agents, Harnesses, and Enterprise Tooling** **Harnesses becoming first-class**: A recurring theme was that model capability is increasingly mediated by the** agent harness**.[@omarsar0](https://x.com/omarsar0/status/2093056965568332236)highlighted** JIT-Agent**, where the model synthesizes a harness over modules for memory, planning, action protocol, and tool orchestration, reporting gains over off-the-shelf agents. Separately,[@dair_ai](https://x.com/dair_ai/status/2093030540807213178)shared work inducing compact**finite-state machines from agent traces**, suggesting behavior topology may be shaped more by deployment scaffolds than by the underlying LLM.** Product releases around agent infra**: Anthropic released a cookbook for connecting** Claude Managed Agents**to** Vercel’s Chat SDK**, giving a unified chat layer with server-side harness, session management, and memory[@ClaudeDevs](https://x.com/ClaudeDevs/status/2092984433649283284). Perplexity added**connectors in Agent API** for**GitHub, Slack, Google Drive, and Datadog**[@perplexitydevs](https://x.com/perplexitydevs/status/2092975514558550102). Cursor announced a workflow to create web apps, store code with Origin, and deploy to Vercel[@cursor_ai](https://x.com/cursor_ai/status/2093077548649570777).**Higher-trust browser automation**: Nous shipped a significant escalation for browser-use agents:** Hermes Agent can now browse as you**, using a managed copy of your** real Chrome profile / logins**[@NousResearch](https://x.com/NousResearch/status/2093063359587348487),[@Teknium](https://x.com/Teknium/status/2093064288877547760). This is a notable usability boost, but it also materially changes the risk surface for cloud agents by collapsing auth friction and making scoped-permission design much more urgent. **Security, Agent Misalignment, and Cyber Defense Coordination** **OpenAI-led cyber defense coalition**: OpenAI published an** open letter**signed by** 116 organizations**including Anthropic, AWS, Google, Microsoft, and Oracle, calling for a global surge in cyber defense against AI-enabled attacks[@OpenAI](https://x.com/OpenAI/status/2093074192636018977), with Sam Altman stressing that “there is not much time to act”[@sama](https://x.com/sama/status/2093060670472241368). Regardless of one’s policy priors, this was one of the day’s clearest cross-industry coordination moves.**Double-blind frontier evals**: Google DeepMind announced a pilot for** double-blind evaluations**of frontier AI, using a secure environment where** neither test prompts nor model weights are revealed**[@GoogleDeepMind](https://x.com/GoogleDeepMind/status/2092961763553677387). For practitioners, the key significance is procedural: a serious attempt to make external evals possible without giving either side full visibility into the other’s assets.**Agent incident analysis continues**: Discussion around the OpenAI/Hugging Face agent incident remained active. Researchers involved in the investigation shared extra details about large transcript sweeps, collaboration patterns among agents, and later swarms apparently building on earlier work[@RyanGreenblatt](https://x.com/RyanGreenblatt/status/2093047632830845016),[@HjalmarWijk](https://x.com/HjalmarWijk/status/2093143101246423436),[@ajeya_cotra](https://x.com/ajeya_cotra/status/2093144336024355104). A separate paper summary from[@omarsar0](https://x.com/omarsar0/status/2093001097346764950)on**EvoMal** warned that shared skill libraries can become**self-poisoning malware propagation channels** for coding agents. Together these point to a maturing realization: multi-agent systems introduce failure modes that are neither classic software bugs nor standard model eval issues. **Top tweets (by engagement)** **Microduck dominates mindshare**: The highest-signal product buzz centered on[@ClementDelangue’s Microduck announcement](https://x.com/ClementDelangue/status/2092931447644442635),[@Thom_Wolf’s technical launch thread](https://x.com/Thom_Wolf/status/2092923071829049592), and follow-up sales milestones from[@Thom_Wolf](https://x.com/Thom_Wolf/status/2093023975173431449).**Cyber defense call gets major traction**: The strongest policy/security engagement came from[@sama](https://x.com/sama/status/2093060670472241368)and[@OpenAI](https://x.com/OpenAI/status/2093074192636018977)on collective cyber defense.**Anthropic’s science push lands**:[@claudeai](https://x.com/claudeai/status/2093059087298601113)announced a** Claude Team plan for scientists**covering** 10,000 researchers**, with free standard seats and** premium seats at $15/month for a year**.** Hermes browser access stands out**:[@NousResearch](https://x.com/NousResearch/status/2093063359587348487)drew substantial engagement for giving agents access to a user’s**real browser profile**, one of the more consequential UX/security tradeoffs in current agent tooling.