Thinkless: Fight against the Growing Resource Demands of AI

In recent months, major tech companies have announced a series of reasoning features in their models. However, the immense resource requirements of these systems quickly became apparent, causing the prices of such subscription services to soar. Researchers at the National University of Singapore (NUS) have developed a new framework called "Thinkless", which could significantly transform how large language models (LLMs) handle reasoning tasks. This innovative approach, created by Gongfan Fang, Xinyin Ma, and Xinchao Wang at the NUS xML Lab, enables AI systems to dynamically choose between simple and complex reasoning strategies—potentially reducing computational costs by up to 90%. The framework addresses a critical inefficiency in current AI reasoning methods and represents a major step toward more resource-efficient AI.

Large language models have demonstrated impressive reasoning abilities using techniques like chain-of-thought (CoT), which facilitates step-by-step logical reasoning. However, while CoT improves performance on complex tasks requiring multi-step reasoning, it also incurs significant computational costs. The extensive reasoning tokens generated during these processes increase latency, memory usage, and overall computational demand.

The core issue identified by the researchers is that not every problem requires complex reasoning. Many queries can be answered with simple responses, yet current models often apply the same elaborate reasoning process regardless of task complexity.

This one-size-fits-all approach leads to substantial computational waste, particularly when the tasks are simple. For example, asking an AI to solve a basic addition problem like "2+2" often triggers the same resource-intensive reasoning pipeline used for complex mathematical proofs.

This inefficiency becomes especially problematic at scale—when LLMs process millions of queries daily, every unnecessary inference step multiplies energy consumption, costs, and environmental impact. Recognizing this challenge, researchers at NUS proposed Thinkless as a solution to improve efficiency.

The Thinkless framework enables LLMs to adaptively select between short and long-form reasoning depending on task complexity and model capabilities. The system uses two control tokens: <short> for concise answers and <think> for detailed reasoning.

This binary control mechanism allows the model to dynamically determine the appropriate reasoning depth for each query.

At the heart of the Thinkless method is the novel Decoupled Group Relative Policy Optimization (DeGRPO) algorithm, which marks a significant improvement over traditional reinforcement learning methods. DeGRPO breaks down the learning objective into two key components:

  • Control token loss – Governs the selection of reasoning mode, helping the model learn when to use extensive reasoning and when to deliver direct answers.

  • Response loss – Improves the accuracy of generated responses, ensuring high performance regardless of the chosen reasoning strategy.

This decoupled structure enables fine-grained control over each objective and stabilizes model training.

The model is trained within a reinforcement learning framework, allowing it to learn from experience which problem types benefit from detailed reasoning and which can be efficiently addressed with minimal computation.

The research team developed several model variants, including a 1.5-billion-parameter reinforcement learning model called "Thinkless-1.5B-RL-DeepScaleR" and a warm-up variant named "Thinkless-1.5B-Warmup."

NUS researchers evaluated Thinkless on multiple benchmarks, including Minerva Algebra, MATH-500, and GSM8K. The results were impressive: the framework reduced reliance on long-chain reasoning by 50–90%, significantly enhancing the computational efficiency of inference in language models.

Crucially, this increase in efficiency was achieved without sacrificing accuracy—a key result, as previous attempts to reduce inference complexity have often led to performance degradation. The decoupled nature of the DeGRPO algorithm allows Thinkless to maintain high performance while substantially lowering the computational burden.

Thinkless is not the only approach targeting the computational demands of AI reasoning. Several other strategies have been developed to tackle similar challenges:

ThinkLess (a training-free method)

Confusingly, another research team led by Gengyang Li has developed a similarly named approach called "ThinkLess." However, this method differs fundamentally from NUS’s Thinkless framework as it does not require model training.

Instead of training the model to select reasoning strategies, this method terminates inference early while preserving output quality. Researchers found that response cues often rely more on termination signals than on prior reasoning steps. ThinkLess exploits this insight by triggering the terminator token earlier to skip redundant reasoning while maintaining knowledge transfer.

Early-Stopping Self-Consistency (ESC)

Proposed by Yiwei Li et al., ESC is a scalable sampling technique designed to reduce the cost of self-consistency (SC) in multi-step reasoning. Traditional SC requires multiple samples at a predefined scale, which increases computation. ESC dynamically adjusts this process and significantly reduces the average number of samples—by 33.8% to 84.2%, depending on the benchmark—while maintaining comparable performance.

Dynamic Chain-of-Thought (D-CoT)

D-CoT implements adaptive reasoning depth and steps to reduce redundancy and delays in reward assignment during extended CoT reasoning. It introduces a state compression mechanism, adaptive inference steps, and an importance-driven pruning strategy during autoregressive decoding. It also features a partial reward estimator to assess the efficiency of inference blocks in real time, constructing a multi-level reasoning structure via macro- and micro-constraint buffers.

As concerns grow about the carbon footprint of AI, these approaches mark a crucial step toward more environmentally sustainable AI systems. Moreover, more efficient reasoning is also key to deploying advanced AI capabilities on devices with limited computational resources.

Share this post
Sovereign AI, secret share sales – what is going on behind the scenes at NVIDIA?
The artificial intelligence industry has experienced unprecedented momentum in recent years, and one of the biggest winners of this wave is undoubtedly NVIDIA. Known for its graphics processors, the company is now not only a favorite among gamers and engineers, but has also become a central player in international technology strategies. Its shares are hitting historic highs on the US stock market, while more and more government cooperation and geopolitical threads are beginning to weave around it. But what does all this tell us about the future, and how well-founded is the current optimism?
GNOME 49 will no longer support X11
Although GNOME is perhaps the most commonly used desktop environment for individual Linux distributions, the developers have decided to make deeper structural changes in GNOME 49, which will affect distribution support.
Facebook's new AI feature quietly opens the door to mass analysis of personal photos
Users who want to share a post on Facebook are greeted with a new warning: a pop-up window asking for permission for “cloud-based processing.” If we approve, the system can access our entire phone photo library—including photos we've never uploaded to the social network. The goal: to generate creative ideas using artificial intelligence, such as collages, themed selections, or stylized versions.
openEuler 24.03-LTS-SP2 is the platform of choice for large enterprises in China
The future of digital infrastructure is increasingly based on operating systems that can meet the stability, innovation and compatibility requirements of different industries. openEuler, China's first community open source operating system, is not just a technology product, but the result of a long-term strategic effort to create an independent and diverse technology ecosystem. The latest major milestone in this development is openEuler 24.03 LTS SP2.
Google Gemini CLI, a powerful offering in the field of AI accessible from the terminal
Google's recently announced Gemini CLI is an open source, command line AI tool that integrates the Gemini 2.5 Pro large language model directly into the terminal. The goal of the initiative is nothing less than to transform natural language commands into real technical workflows, in an environment that has already been synonymous with efficiency for many.
Satya Nadella's thoughts on the role, future, and responsibility of artificial intelligence
Rapid change is not uncommon in the world of technology, but rarely does it affect so many sectors at once as today's artificial intelligence (AI) revolution. In an interview with Y Combinator, Satya Nadella, CEO of Microsoft, not only assessed technological developments, but also placed the development of AI in a broader social and economic context. His approach is restrained, calm, and purposeful: AI is not a mystical entity, but a tool that must be properly applied and interpreted.

Linux distribution updates released in the last few days