Editorial illustration for Alibaba's Qwen team adds method that lengthens AI answers, prompting reasoning
Alibaba's Qwen AI Breakthrough: Longer, Smarter Answers
Alibaba's Qwen team adds method that lengthens AI answers, prompting reasoning
Forget subtle tweaks. Alibaba's Qwen team found a brutally simple lever: force the AI to write more. A new paper details this mechanical lengthening, compressing the entire distribution of answers upward.
The shortest output from the new system now exceeds the longest from the old. This procedural mandate, they discovered, makes the model reason longer. It starts fact-checking its own logic.
Reinforcement learning hits a wall with reasoning models because every token gets the same reward. A new algorithm from Alibaba's Qwen team fixes this by weighting each step based on how much it shapes what comes next, doubling the length of thought processes in the process.
That's the breakthrough—the shift from a clean chain of thought to a relentless internal audit. The core trick, per the paper, was making verbosity a training objective. Rewarded for length, the model learned to interrogate its own plans.
It loops back to correct itself. This isn't about word count. It’s instilling doubt.
The result is a model whose answers, long or short, have passed a procedural review. For now, it’s a brute-force research finding: sometimes, the only way to get a better answer is to demand a longer one.
Related Reading
Research & Benchmarks Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks Databricks finds multi-step agents beat single-turn RAG by 21% to 38% on STaRK
Research & Benchmarks Nvidia's DLSS 4.5 beta adds 6x Multi Frame Generation for RTX 50 GPUs
From Archives Open models cross threshold; frontier models show per‑category correctness
From Archives Batch Mode VC-6 and NVIDIA Nsight Speed Up Vision AI Pipelines
Common Questions Answered
How does Alibaba's Qwen team's new algorithm change language model response generation?
The new algorithm nudges language models to produce longer, more reflective replies by shifting the distribution of answer lengths upward. This approach moves beyond terse responses, encouraging models to develop more nuanced reasoning and self-verification processes.
What are the four distinct training phases described by the Qwen team's research?
The research outlines four training phases where the model evolves from generating shallow planning templates to more sophisticated reasoning. In the early stages, the model produces initial outlines with limited depth, gradually developing more complex reasoning chains and self-checking mechanisms.
How does the new algorithm change token weighting in language models?
The Qwen team's algorithm reweights tokens according to their downstream influence, moving away from the uniform token treatment in earlier models. This approach allows the model to develop more extensive reasoning chains and compare alternative solutions more effectively.
Further Reading
- Alibaba's new Qwen reasoning AI model sets open-source records — Artificial Intelligence News
- QwenLong-L1.5: Post-Training Recipe for Long-Context Reasoning ... — arXiv
- Qwen AI Models 2025: Alibaba's Advanced Multilingual AI Family for ... — Local AI Zone
