ChartVerse: Scaling Chart Reasoning via Reliable Programmatic Synthesis from Scratch

2026-01-20

AI Summary

This work addresses the limitations of open-source vision-language models in chart reasoning tasks, which stem from a scarcity of high-quality training data—existing synthetic datasets are often overly simplistic and prone to hallucinated question-answer pairs. To overcome this, the authors propose ChartVerse, a framework that programmatically generates charts of high visual and semantic complexity and employs an answer-anchored inverse question synthesis mechanism coupled with rigorous consistency verification to produce reliable reasoning data. They introduce a complexity-aware generator guided by a novel Rollout Posterior Entropy–based metric, along with failure-rate filtering and Chain-of-Thought distillation to enhance reasoning depth. Using this approach, they construct the ChartVerse-SFT-600K and ChartVerse-RL-40K datasets, enabling the trained ChartVerse-8B model to outperform the teacher model Qwen3-VL-30B-A3B-Thinking and match the performance of the stronger Qwen3-VL-32B-Thinking.

Technology Category

• Computer Vision: Large Vision Models
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Chart reasoning is a critical capability for Vision Language Models (VLMs). However, the development of open-source models is severely hindered by the lack of high-quality training data. Existing datasets suffer from a dual challenge: synthetic charts are often simplistic and repetitive, while the associated QA pairs are prone to hallucinations and lack the reasoning depth required for complex tasks. To bridge this gap, we propose ChartVerse, a scalable framework designed to synthesize complex charts and reliable reasoning data from scratch. (1) To address the bottleneck of simple patterns, we first introduce Rollout Posterior Entropy (RPE), a novel metric that quantifies chart complexity. Guided by RPE, we develop complexity-aware chart coder to autonomously synthesize diverse, high-complexity charts via executable programs. (2) To guarantee reasoning rigor, we develop truth-anchored inverse QA synthesis. Diverging from standard generation, we adopt an answer-first paradigm: we extract deterministic answers directly from the source code, generate questions conditional on these anchors, and enforce strict consistency verification. To further elevate difficulty and reasoning depth, we filter samples based on model fail-rate and distill high-quality Chain-of-Thought (CoT) reasoning. We curate ChartVerse-SFT-600K and ChartVerse-RL-40K using Qwen3-VL-30B-A3B-Thinking as the teacher. Experimental results demonstrate that ChartVerse-8B achieves state-of-the-art performance, notably surpassing its teacher and rivaling the stronger Qwen3-VL-32B-Thinking.

Problem

• chart reasoning
• training data scarcity
• hallucination
• reasoning depth
• synthetic charts

Innovation

• Chart Reasoning
• Programmatic Synthesis
• Rollout Posterior Entropy
• Truth-Anchored QA
• Chain-of-Thought Distillation