Zhipu's launch of GLM-5.1 and Cognition's release of SWE-1.6 were the big highlights with respect updates to agentic codel models. OpenAI's $100/month ChatGPT Pro plan with a 5x Codex usage definitely makes it quite attractive. Elon Musk teases multiple large xAI models (up to 10T parameters) in training
Readers who think others in their family, friends and acquaintances who are curious in knowing more about rapidly evolving AI tools/services/use cases and would benefit from being subscribers of this weekly newsletter are encouraged to share this publication link to them and invite them to subscribe
The following messages were posted on the 'All Things AI' Telegram group from Sunday 5th Apr 2026 to Saturday 11th Apr 2026
From RC who previously built Kimi CLI for Moonshot AI comes an early glimpse of Slock of what he terms an agent-human collaboration platform for modern builders and teams. Pricing not yet defined, he says that it will be free for the next 2 weeks. Hermes Agent support and Social Auth login in progress
Came across this on the OpenAI Codex rate card page
Note: As of April 2, 2026, we’ve updated Codex pricing to align with API token usage, instead of per-message pricing. This change is applicable to new and existing ChatGPT Business and new ChatGPT Enterprise plans - please refer to the new rate card in the section below for details.
Customers on existing Plus, Pro and Enterprise/Edu plans should continue to use the legacy rate card. We’ll migrate you to the new rates in the upcoming weeks.
All customers should carefully review both rate cards to understand the pricing changes, and how they apply to your plan.
https://help.openai.com/en/articles/20001106-codex-rate-card
Actress Milla Jovovich famous for portraying Alice in the Resident Evil series in her instagram story shares about her architecting MemPalace, a new open-source AI memory storage and retrival system with the highest LongMemEval score ever published
https://www.instagram.com/reel/DWzNnqwD2Lu/
https://linktr.ee/mempalace
https://github.com/milla-jovovich/mempalace
Recently came across this profile of Accio which I had mentioned earlier in this group via MIT Technology Review
====
A look at Alibaba's Accio, an AI sourcing tool that helps small online sellers connect with manufacturers, including in China; it exceeded 10M MAUs in March
AI is changing how small online sellers decide what to make
Entrepreneurs based in the US are using tools like Alibaba's Accio to compress weeks of product research and supplier hunting into a single chat.
Z.ai Launches GLM-5.1, Top Open AI for Coding and Long Tasks
Z.ai released GLM-5.1, a 754-billion-parameter open-weights model excelling in coding and extended agent tasks, topping open-source benchmarks like SWE-Bench Pro at 58.4% and Terminal-Bench 2.0 at 66.5%. It outperformed GPT-5.4 and Claude Opus on self-reported scores by autonomously handling demos such as building a complete desktop with 50+ apps or boosting a vector database 6x via thousands of tool calls. Weights are freely available on Hugging Face under MIT license, with quantized versions for local runs and API access through multiple platforms,
agents could do about 20 steps by the end of last year. glm-5.1 can do 1,700 rn. autonomous work time may be the most important curve after scaling laws. glm-5.1 will be the first point on that curve that the open-source community can verify with their own
Anthropic Launches Project Glasswing to Bolster Software Security with AI
Project Glasswing gives vetted partners early access to Claude Mythos Preview, an advanced AI that excels at finding and exploiting vulnerabilities, scoring far higher than previous models on key benchmarks. It has already uncovered thousands of critical flaws in major operating systems, browsers, FFmpeg, and the Linux kernel, leading to patches. Members like Microsoft, AWS, Apple, Google, and Cisco are strengthening their codebases, with Anthropic providing $100 million in compute credits
It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans.
Project Glasswing: Securing critical software for the AI era
A new initiative to secure the world’s most critical software and give defenders a durable advantage in the coming AI-driven era of cybersecurity.
via Ben Jammin
Good timing on this one
Run Gemma 4 locally with OpenClaw in 3 steps
Cognition Releases SWE-1.6 AI Model in Windsurf with Improved Intelligence and Speed
Cognition released SWE-1.6, an advanced AI model matching SWE-Bench Pro performance while enhancing behavioral aspects and user experience through post-training optimizations that reduce assistant turns by 40%. The model is available in Windsurf with a free tier at 200 tokens per second and a fast tier at 950 tokens per second.
It’s available today in Windsurf in two modes: free tier (200 tok/s) and fast tier (950 tok/s).
Consolidating some opencode news. PS, Parallel was founded by Parag Agarwal, Twitter's former CEO
Give your agent more reliable, accurate, and efficient web search tools: npmjs.com/package/@paral…
Amazon S3 Files gives AI agents a native file system workspace, ending the object-file split that breaks multi-agent pipelines
S3 Files and the changing face of S3
Andy Warfield writes about the hard-won lessons dealing with data friction that lead to S3 Files
The first and only cloud object store with fully-featured, high-performance file system access.
Learn more here. go.aws/4tw17Zg
Only recently did I find out that Ollama offered a subscription which offered cloud access to open-source models
Ollama's cloud is one of the best places to run OpenClaw. $20 plan is enough for most day to day OpenClaw usage with open models!
To make the switch, all you need is to open the terminal and type:
ollama launch openclaw
Choose a model:
kimi-k2.5:cloud
Max subscription supports 10 models at once.
Pricing · Ollama
Get up and running with large language models.
Looks like Minimax is betting on Hermes from Nousresearch giving it more distribution
MiniMax models are already some of the most-used in Hermes Agent. If you haven’t tried MiniMax M2.7 in Hermes Agent, try it today in the Nous Portal!
via Robby Yung
If you're building with AI agents, this is for you.
I've been using this for 2 months. It's totally changed how I do my job.
You install it in your coding agent. Cursor, Claude Code, Codex, whatever you use. Now that agent has access
Anthropic Launches Claude Managed Agents in Public Beta
The public beta offers APIs with a pre-built agent harness that manages secure sandboxes, stateful sessions, error recovery, and scaling for long tasks. Developers define tasks, tools, and instructions while the system runs code, browses the web, and integrates with tools like MCP—all viewable in the Claude Console. Companies like Notion, Rakuten, Asana, and Sentry are already building with it, from department agents to AI teammates. Early users report up to 10x faster development and higher success on tough tasks
The screenshot tells the full story. That's a production fleet dashboard. 8 agents running. 247 completed tasks. Active status. MCP-connected to HubSpot, pulling deals, generating proposals,
via Cham Ho
I would be interested to see if anyone has recommendations for AI code review tools? Am testing a few out at the moment so would welcome any additional recommendations
via Brent Maxwell
I recommend Devin: https://app.devin.ai/review. We've been using Cursor's BugBot for a year, and I've tried a few other options (including rolling our own), but Devin is EXCELLENT. We implemented it about a month ago, and it's basically a replacement for GitHub PR reviews as well as Bugbot.
In our experience, a lot of these PR agents pick up individual bits of code that needs work, but Devin aggregates findings across multiple files and groups them into functional problems and improvements. The UI is very seamless and rich.
Sounds like a paid promotion: it's not. I'm just excited because this has solved some shortcomings that have been annoying me for about a year. Just give it a free test by pointing it at any open source repo. Literally 10m to see whether you like it or not.
Came across this Github repo with what I consider decent content on Hermes.
A 75+ page book published both in English and Chinese
GitHub - alchaincyf/hermes-agent-orange-book: Hermes Agent 从入门到精通 · 橙皮书系列 · Nous Research 开源 AI Agent 框架实战指南
Hermes Agent 从入门到精通 · 橙皮书系列 · Nous Research 开源 AI Agent 框架实战指南 - alchaincyf/hermes-agent-orange-book
OpenAI introduces ChatGPT Pro $100 tier with 5X usage limits for Codex compared to Plus
Expectations high for these upcoming models from xAI, hopefully they are competitive in agentic coding and also offer good price and are well supported by all the harnesses. June-July is going to get very interesting.
via Ben Jammin
This is really interesting
“every app built on Anything is now buyable, remixable, sellable”
via Ben Jammin
MMX-CLI gives every Agent seven new senses — image, video, voice, music, vision, search, conversation
Below is my personal website which aggregates links to many of my socials as well as the various content and community that I curate. Feel free to share this link to others who you think may find this content/community useful to them
The cover image of this newsletter via generated via the Google Nano Banana 2 model within the Freepik tool via the following prompt
A mesmerizing land art masterpiece, featuring an enormous, intricate design of a mythical sparrow creature crafted meticulously using arich palette of stones. The creature's stunning details are made by the contrasting colors and textures of the stones. The serene beach in the background. The diverse array of stones, skillfully arranged to create this magnificent work, evoke a sense of awe and appreciation for the beauty of nature and the artist's talent.