ICANEWS

Lightning Weave: Composing Capabilities for Improved Reasoning Model Accuracy-Efficiency Frontier

arXiv CS · · 2 min read · Engineering & Technology

Read research and analysis on Lightning Weave: Composing Capabilities for Improved Reasoning Model Accuracy-Efficiency Frontier published by ICANEWS, a global research journal for emerging researchers.

Key Takeaways

  • Lightning Weave extracts and composes independently learned capabilities into a single student model via on-policy distillation.
  • The framework combines aligned log-ratio shifts at shared student token states and uses Tilted-Target DOPD for stable learning targets.
  • On Qwen3.5-4B, Lightning Weave raised HMMT 2025 accuracy from 59.2% to 64.0% with 10.7% fewer response tokens.
  • On Qwen3.5-4B, LiveCodeBench v5 accuracy increased from 41.7% to 54.2% with 9.6% fewer response tokens.
  • Adjusting anchor signal strengths yields a strong empirical accuracy-efficiency Pareto frontier.

Why This Matters

The framework offers a practical method to improve the accuracy-efficiency frontier in reasoning models, indicating a pathway for more effective model deployment. Its ability to achieve higher accuracy with fewer response tokens could lead to more resource-efficient reasoning systems.

Overview

Lightning Weave is introduced as a post-training framework designed to enhance the accuracy and inference efficiency of reasoning models by composing independently learned capabilities. The framework addresses the challenge of jointly improving these two objectives, which can favor distinct reasoning behaviors.

Research Context

Efficient reasoning aims to improve the accuracy-efficiency frontier. Independently post-trained models are noted to possess distinct strengths in terms of accuracy and efficiency. The framework's core premise involves leveraging these pre-existing specialist capabilities.

Approach

Lightning Weave functions by extracting and composing independently learned capabilities within a single student model. This process occurs through on-policy distillation. Each acquired capability is represented as the policy shift from the model prior to post-training to the resulting specialist model. The framework combines aligned log-ratio shifts at shared student token states. To convert cached signals into a stable learning target, Lightning Weave utilizes Tilted-Target DOPD. A key operational aspect is that each anchor pair scores cached trajectories only once, which enables subsequent student training without requiring multiple live anchor models to be served concurrently.

Findings

Across diverse student models and benchmarks in mathematics and code, Lightning Weave demonstrated substantial improvements over base students and achieved a state-of-the-art accuracy-efficiency frontier. Specific results include:

  • On Qwen3.5-4B, Lightning Weave raised HMMT 2025 accuracy from 59.2% to 64.0%. This improvement was accompanied by a 10.7% reduction in response tokens.
  • For the same model, LiveCodeBench v5 accuracy increased from 41.7% to 54.2%. This was achieved with 9.6% fewer response tokens.

Empirical observations indicated that adjusting the relative strengths of the anchor signals yields a strong empirical accuracy-efficiency Pareto frontier.

Why This Matters

These results establish Lightning Weave as a new practical route to efficient reasoning through the composition of capabilities. The framework's ability to improve both accuracy and reduce response tokens suggests potential for more efficient deployment and operation of reasoning models.

Research Information

Institution
arXiv
Original Study
View Publication
Source
arXiv CS

About ICANEWS

ICANEWS is a global research journal for emerging researchers, publishing student and emerging researcher work across all fields.