AI News Sep 18, 2026
By Frontier Editorial •
Key Takeaways
- World Labs releases Atlas AI model to generate 3D worlds from images: World Labs has introduced Atlas, an AI model that can generate controllable 3D environments from a small numbe…
- Zhipu launches GLM-5.3-FlashX with 200 tokens/s inference speed: Zhipu AI has launched GLM-5.3-FlashX, claiming inference speeds of nearly 200 tokens per second on approximately 10…
- Researchers hack OpenAI's GitHub monorepo via bug bounty program: Independent security researchers participating in OpenAI's bug bounty program successfully hacked the company, gai…
- Magentic: 18M (Series A)
- 1. Open your Claude configuration file located at ~/.claude/settings.json.
What are the top AI breakthroughs?
This Sep 18, 2026 covers 10 curated AI news items spanning technology, research, and product developments. Zhipu launches GLM-5.3-FlashX with 200 tokens/s inference speed: Zhipu AI has launched GLM-5.3-FlashX, claiming inferenc...
Zhipu launches GLM-5.3-FlashX with 200 tokens/s inference speed: Zhipu AI has launched GLM-5.3-Flash…
Zhipu launches GLM-5.3-FlashX with 200 tokens/s inference speed: Zhipu AI has launched GLM-5.3-FlashX, claiming inference speeds of nearly 200 tokens per second on approximately 100,000 domestic accelerators. The base Flash model is a 320-billion parameter Mixture of Experts model with 18 billion active parameters, optimized by an Infra Agent.
Amap launches ABot-Earth 0.7 to generate 3D cities 1,000 times faster: Amap has released ABot-Earth …
Amap launches ABot-Earth 0.7 to generate 3D cities 1,000 times faster: Amap has released ABot-Earth 0.7, a 3D-native urban world model integrated with Flying Street View 2.0. The model can generate kilometer-scale 3DGS cities on a single consumer GPU in approximately 10 minutes using satellite or text input. This process is reportedly about 1,000 times faster than traditional pipelines.
Volcengine launches Doubao-Seed-2.1-pro with multimodal coding: Volcengine has fully released the Do…
Volcengine launches Doubao-Seed-2.1-pro with multimodal coding: Volcengine has fully released the Doubao-Seed-2.1-pro 0915 model featuring multimodal coding capabilities. The model demonstrated rebuilding a 280,000-line Java ERP from screen recordings and sketches, achieved an 83% mergeable rate on Luanti repo fixes, and reduced image and video inference token costs by over 30%.
PrismML compresses Alibaba's Qwen3.8 to 5.9 GB for smartphones: PrismML has released Bonsai 2 27B, a…
PrismML compresses Alibaba's Qwen3.8 to 5.9 GB for smartphones: PrismML has released Bonsai 2 27B, a model that compresses Alibaba's Qwen3.8 27B down to 5.9 GB. This compression makes the model small enough to run on smartphones while retaining 98.2% of Qwen's original benchmark scores.
Anthropic updates Claude Code with parallel agent workflows in beta: Anthropic has rebuilt Projects …
Anthropic updates Claude Code with parallel agent workflows in beta: Anthropic has rebuilt Projects in Claude Code to introduce parallel agent workflows. A coordinator now splits tasks across parallel cloud threads that share a common memory, independently opening pull requests and running tests. The beta is available to select Pro and Max subscribers.
Researchers hack OpenAI's GitHub monorepo via bug bounty program: Independent security researchers p…
Researchers hack OpenAI's GitHub monorepo via bug bounty program: Independent security researchers participating in OpenAI's bug bounty program successfully hacked the company, gaining access to its internal "monorepo" on GitHub. The team executed the breach using cybersecurity versions of Opus 4.8 and Opus 5.
OpenAI launches Astra for Law powered by GPT-6 Astra for select firms: OpenAI has launched Astra for…
OpenAI launches Astra for Law powered by GPT-6 Astra for select firms: OpenAI has launched Astra for Law, a new AI foundation designed for legal analysis and writing. The tool combines the GPT-6 Astra model with a specialized legal search index and tailored instructions. It is initially available to select law firms.
Huawei plans Q1 2027 launch of Ascend 960DT AI chip to rival Nvidia: Huawei is accelerating the laun…
Huawei plans Q1 2027 launch of Ascend 960DT AI chip to rival Nvidia: Huawei is accelerating the launch of its next-generation Ascend 960DT AI chip, targeting a release in the first quarter of 2027. The move is part of the company's push to compete with Nvidia and narrow China's AI computing gap with the United States.
World Labs releases Atlas AI model to generate 3D worlds from images: World Labs has introduced Atla…
World Labs releases Atlas AI model to generate 3D worlds from images: World Labs has introduced Atlas, an AI model that can generate controllable 3D environments from a small number of images. The system is capable of filling in areas not captured by the original camera.
Researchers introduce MAGS framework to guarantee safety for agentic outputs: Researchers have intro…
Researchers introduce MAGS framework to guarantee safety for agentic outputs: Researchers have introduced MAGS, a unified multi-agent framework designed to generate executable programs with formal safety guarantees. The framework utilizes Dafny as a verification-aware intermediate representation to address the limitations of traditional manual specification and proof engineering.
What are the latest AI investment signals?
Latest AI investment signals: 2 funding rounds, 0 market updates, and 0 M&A transactions.
Primary Market – Funding Rounds
Secondary Market – Market Updates
No secondary market data.
M&A – Mergers & Acquisitions
No M&A data.
What are practical AI tips this week?
1 practical AI tips curated from Reddit communities and expert blogs. Claude Code users can save costs by extending sub-agent prompt cache TTL to one hour...
Claude Code
Claude Code users can save costs by extending sub-agent prompt cache TTL to one hour
1. Open your Claude configuration file located at ~/.claude/settings.json. 2. Add or update the setting "subagentPromptCacheTtl": "1h" in the JSON file. 3. Save the file to extend the sub-agent prompt cache lifetime from 5 minutes to 1 hour.