3/3 Across 4 benchmarks and 3 LLM backbones, PoS achieves the best overall results among evaluated methods in all 12 settings.
Paper and code are out—feedback welcome!
📄 https://t.co/7p0Q0ZMpDB
💻 https://t.co/YRUA9eyKqc
1/3 Excited to share PoS (Progression of States)! 🚀
Long-horizon agents need to track a changing world. But they can keep reasoning and calling tools without making progress.
We introduce PoS to detect and recover from this failure mode: Belief Trapping. 🧵
Alibaba's PoS: Beyond memory for long-horizon agents
An inference-time framework that maintains explicit belief states, detects when agents stall (Belief Trapping), and recovers with targeted constraints. Achieves SOTA across 4 benchmarks, with gains up to 37.89%.
2/3 PoS maintains explicit belief states: what the world looks like, what remains unknown, and what still needs to be done.
It validates updates, diagnoses stagnation, cycles, or drift, and guides recovery based on what's blocking progress.
No extra training.
GPT6.1 Sol works well as a personal agent, but it’s extremely slow! And I don’t like dots because they don’t show me how the task is progressing. I can’t tell whether it’s actually doing what I asked it to do.
Introducing Graph of States (GoS), a neuro-symbolic framework for LLM-based abductive reasoning and multi-agent diagnosis. It enables evidence-guided search over explicit belief states. Accepted at ICML 2026.
arxiv:https://t.co/UFR5TtmMcW
Github:https://t.co/AxjnViO2lz
Introducing OpsAgent, a lightweight lightweight, self-evolving multi-agent system for incident management in microservices. Accepted at ASE 2026.
arXiv:https://t.co/0qOQFxNUZi
GitHub:https://t.co/13Twb4TZ1O
📢Meet Qwen3.8-Max — our most capable model to date.
Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉
Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters:
- Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub:https://t.co/iVHZWQoeSo
- Real work, real results: Production-quality deliverables across hundreds of professions.
- Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy.
- Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction.
💰Pricing:
Input: $2.0 / M tokens
Output: $6.0 / M tokens
Implicit Caching: $0.25 / M tokens
Start building with Qwen3.8-Max! 🚀
📖 Blog: https://t.co/iwjmQxLBof
✅ Qwen Studio: https://t.co/4V2pFvDovG
⚡ API: https://t.co/gAGqaLQGbN
It is a hard and sad decision. I shared this message with folks at Thinky. Thank you all for the time together♥️ Just as the last sentence in my message: The future worth building is human.