arxiv Moonshot AI (arXiv) · 440d ago

Alita enables scalable agentic reasoning with minimal predefinition and maximal self-evolution

from English

Researchers introduce Alita, a generalist agent designed to perform complex tasks through minimal predefinition and maximal self-evolution, avoiding the heavy reliance on manually predefined tools found in many existing frameworks.

Alita uses only one component for direct problem-solving to enhance adaptability and generalization across domains.
The system autonomously constructs, refines, and reuses external capabilities by generating task-related model context protocols (MCPs) from open source.
On the GAIA benchmark validation dataset, Alita achieves 75.15% pass@1 and 87.27% pass@3 accuracy.
It scores 74.00% pass@1 on Mathvista and 52.00% pass@1 on PathVQA, outperforming more complex agent systems.

This approach allows Alita to scale agentic reasoning effectively while maintaining a simpler architecture than previous elaborate tool-based approaches.

Importance 1/3 New harness with differentiators Moonshot AI (arXiv) AI agents Research paper

Benchmarks

Benchmark	Model	Score
GAIA	Alita	75.15%
MathVista	Alita	74%

Read original

arxiv arXiv cs.AI · 3d ago

Argus: A General-Purpose Agentic Runtime for Long-Horizon Reasoning

Researchers present Argus, a persistent, self-evolving agentic runtime designed for long-horizon reasoning that separates stable user intent from operational objectives. The system utilizes Manager, Planner, Engineer, and Reviewer roles to execute bounded missions over durable project state, allowing for autonomous execution between operator-owned escalation points.

media MarkTechPost · 3d ago

Microsoft's SkillOpt enables agent skill transfer between Codex and Claude Code

Microsoft, Shanghai Jiao Tong University, Tongji University, and Fudan University researchers developed SkillOpt, a text-space optimizer that trains natural-language skill documents while keeping the target model frozen. The system uses an optimizer model to propose bounded edits based on scored rollouts, exporting the result as a single `best_skill.md` file.

arxiv arXiv cs.LG · 5d ago · 6 views

Qwen releases Qwen-CUA, a 397B-A17B native computer-use agent

Qwen introduces Qwen-CUA, a native computer-use agent built on a 397B-A17B mixture-of-experts backbone that operates via screenshots and input events without relying on DOM trees or accessibility metadata. The system utilizes a scaffold to maintain up to 20 active screenshots and was trained using a cloud rollout fleet with nearly 100,000 vCPUs across approximately 40,000 verifiable tasks.

arxiv arXiv cs.CL · 5d ago · 3 views

Qwen releases Qwen-CUA, a 397B-A17B native computer-use agent

Researchers have introduced Qwen-CUA, a native computer-use agent built on a 397B-A17B Qwen mixture-of-experts backbone. The system operates by observing screenshots and acting through keyboard and mouse events, without relying on DOM trees or accessibility metadata.

lab OpenAI News · 8d ago · 16 views

Astra model solves ten open problems in mathematics and theoretical computer science

OpenAI's internal Astra model has generated solutions to ten long-standing open problems across mathematics and theoretical computer science, with the arguments formalized in Lean. These results span high-dimensional geometry, coding theory, group theory, and quantum complexity, addressing questions that had seen no progress for at least a decade.