This NEW AI AGENT is INSANE! 🤯
Ornith 1.0, a new open-source AI agent from Deep Reinforce, has achieved a score of 82.4 on SWE-Bench, surpassing Claude Opus 4.7. The model introduces self-scaffolding reinforcement learning, allowing it to build its own problem-solving framework without human-built instructions, and is available in four versions ranging from 9B to 397B parameters.
Summary
A new open-source AI coding agent called Ornith 1.0 has been released by Deep Reinforce under an MIT license. According to the transcript, it has achieved a verified SWE-Bench score of 82.4, beating Claude Opus 4.7's performance. The most significant innovation highlighted is its self-scaffolding capability—unlike most existing AI coding agents that require human-built scaffolds (instructions for tool usage, failure retry logic, and memory organization), Ornith 1.0 generates its own scaffold dynamically as it solves problems, with its process evolving in real-time.
The model is available in four different sizes to accommodate various computational resources. The 9B parameter version is lightweight enough to run on a laptop, the 35B version operates on hardware that most development teams already possess, and the 397B flagship model is the version achieving the benchmark-beating performance. All versions are released as free, open-source software under the MIT license, allowing users to deploy, fine-tune, and build commercial products on top of the technology without licensing restrictions.
Key Insights
- Ornith 1.0 achieves a SWE-Bench score of 82.4, surpassing Claude Opus 4.7's performance
- Ornith 1.0 introduces self-scaffolding reinforcement learning, enabling the model to build its own scaffold while solving problems rather than relying on human-built instructions
- Most AI coding agents require human-built scaffolds that specify how to use tools, handle retry failures, and organize memory
- Ornith 1.0 is available in four versions: a 9B model for laptops, a 35B model for existing team hardware, and a 397B flagship model
- Ornith 1.0 is released as free, MIT-licensed open-source software that users can deploy, fine-tune, and commercialize
Topics
Transcript
[0:00] This new AI agent is insane. A free open-source AI just beat Claude Opus 4.7. Ornith 1.0 just dropped from Deep Reinforce. Free open-source, MIT licensed, on SWE-Bench verified. It scores 82.4, beating Claude Opus 4.7. And it does something no other open-source coding AI has done before. Most AI coding agents need a human-built scaffold, instructions that tell the model how to use tools, retry failures, and organize memory. Ornith 1.0 builds its own scaffold while solving your problem. Process evolves as it works. [0:31] That's called self-scaffolding reinforcement learning. Four versions available. The 9B runs on a laptop. The 35B runs on hardware most teams already own. The 397B flagship is the one beating Claude. All free,…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
Google AI Studio Just Got A Huge Security Upgrade
Google is reportedly developing a new "Security Check" mode for AI Studio that would apply secure coding practices to AI-generated applications. The feature aims to automate security verification during app development, addressing common vulnerabilities like exposed keys, weak authentication, and data access issues before deployment.
New Gemini Skills Update Is WILD!
Google has launched Gemini Skills, a new feature allowing users to save custom instructions once and reuse them repeatedly by typing a slash command, replacing the discontinued Gems feature. Skills can include reference files, be combined together, and are built on an open file format (skill.md) that allows portability across different AI tools.
NEW Codex CLI Update Is ABSURD!
OpenAI's new Codex CLI update transforms terminal-based programming assistance by enabling multiple AI agents to work simultaneously on different tasks, visible in a full-screen control center interface. The update includes features like Agent View, work trees, voice control, and parallel task execution, fundamentally changing how AI assistance operates from sequential Q&A to autonomous team-based work management.
NEW Hermes Cloud + Agent OS is Crazy! 🤯
Hermes Cloud combined with Agent OS enables users to deploy AI agents in the cloud 24/7 using free models without API keys or local setup complexity. The system provides persistent memory, sandbox isolation, and integration with custom workflows like content research, competitor analysis, and news monitoring.
Manus 2.0 Just Changed AI Agents Forever
Manas 2.0 represents a fundamental architectural overhaul that enables users to create games, websites, videos, and automations through simple text requests, with the key innovation being editable components rather than fixed outputs. The platform uses a new Cascade engine that reduces token usage by 23% and improves task speed by 28%, while maintaining all project elements in one connected workspace.