Skip to content

Lightning Weave: Improving the Accuracy-Efficiency Frontier of Reasoning Models through Capability Composition

Sep 2026 · 0 citations · 66 references
Computer Science

TL;DR

Lightning Weave is introduced, a post-training framework that extracts and composes independently learned capabilities in a single student through on-policy distillation and achieves a state-of-the-art accuracy-efficiency frontier.

Abstract

A core goal of efficient reasoning is to improve the accuracy-efficiency frontier. However, jointly improving reasoning accuracy and inference efficiency can be challenging, as the two objectives can favor different reasoning behaviors. Independently post-trained models already offer distinct strengths in accuracy and efficiency. We introduce Lightning Weave, a post-training framework that extracts and composes these independently learned capabilities in a single student through on-policy distillation. Each acquired capability is represented by the policy shift from the model before post-training to the resulting specialist. Lightning Weave combines aligned log-ratio shifts at shared student token states and uses Tilted-Target DOPD to convert the cached signals into a stable learning target. Each anchor pair scores the cached trajectories once, enabling subsequent student training without serving multiple live anchor models concurrently. Across diverse student models and benchmarks in mathematics and code, Lightning Weave substantially improves upon the base students and achieves a state-of-the-art accuracy-efficiency frontier. On Qwen3.5-4B, it raises HMMT 2025 accuracy from 59.2% to 64.0% with 10.7% fewer response tokens, and LiveCodeBench v5 accuracy from 41.7% to 54.2% with 9.6% fewer response tokens. Adjusting the relative strengths of the anchor signals yields a strong empirical accuracy-efficiency Pareto frontier. These results establish Lightning Weave as a new practical route to efficient reasoning through capability composition. Code is released at https://github.com/jet-ai-projects/Lightning-Weave.

View source

Similar papers

#artificial intelligence Preprint Sep 2026

FusionEval: Benchmarking Model Fusion for Large Language Models

Model fusion aims to integrate the capabilities of multi-expert models into a single model. Existing evaluations offer limited coverage of reasoning with long chains of thought and tool use, while strong average performance can mask substantial losses in individual expert capabilities. We introduce FusionEval, a benchm...

Shuo Cai, Yang-Gan Gu, Yuhang Liu et al. · 0 citations

Frontier Learning: Training LLM Reasoners at the Edge of Capability

Across several reasoning tasks and model families, the proposed frontier learning approach consistently achieves higher relative gains over fixed-pool baselines, demonstrating that effective post-training requires not only selecting useful problems, but continually generating them at the edge of capability.

Robin Faro, S. Ramesh, Ilija Bogunovic et al. · 0 citations
Preprint Oct 2026

SRA: Selective Residual Adaptation for LLM Reasoning Enhancement

Knowledge distillation transfers reasoning capabilities from large language models to lightweight students. However, distilled small models still struggle with multi-step reasoning. We observe that the distribution divergences between teacher and student models along reasoning paths are highly sparse, and target tokens...

Zhen-Xing Dou, Kun-Hua He, Ming-Yang Yi et al. · 0 citations
#machine learning Preprint Oct 2026

When Does Longer Reasoning Help? Predicting Mathematical Reasoning Through Discovery and Execution

Test-time compute can improve mathematical reasoning, but can short-budget runs predict how mathematical reasoning scales with additional compute? We introduce a Discovery--Execution (DE) framework that predicts the aggregate held-out scaling curves through a convolution of strategy discovery and conditional execution....

Adib Hasan, Lay Jain, Thanic Nur Samin · 0 citations
#artificial intelligence Preprint Oct 2026

From Solver Trajectories to Teaching Trajectories: Cognition-Aligned Reasoning Distillation

CARD is an RL-based framework that transforms solver trajectories into teaching trajectories aligned with the student's cognition that outperforms distillation from raw teacher trajectories in in-domain settings while achieving substantial gains in out-of-distribution performance.

Ya-Shuo Luo, Qian-Ren Mao, Zi-Qi Qin et al. · 0 citations

Capable yet Parsimonious: Extracting and Characterizing Hidden Chain-of-Thought in Frontier Models

It is found that Astra exhibits token-efficient directed reasoning, selecting a correct trajectory earlier, while resolving elementary steps internally and externalizing only crucial reasoning, which provides a behavioral lens on frontier-model reasoning beyond benchmark scores.

Xiao-Yu Luo, Tao Ren, Wen-Rui Yu et al. · 2 citations

Related blog posts

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.