VTBench: Evaluating Visual Tokenizers for Autoregressive Image Generation

TL;DR

VTBench is a comprehensive benchmark for evaluating visual tokenizers in autoregressive image generation, revealing that continuous VAEs outperform discrete VTs in preserving image details and structure.

Contribution

The paper introduces VTBench, a new benchmark for systematically assessing visual tokenizers across multiple tasks, highlighting the superiority of continuous VAEs over discrete VTs.

Findings

Continuous VAEs outperform discrete VTs in image reconstruction.

Discrete VTs often lose fine details and text in images.

GPT-4o shows potential as an autoregressive image generator.

Abstract

Autoregressive (AR) models have recently shown strong performance in image generation, where a critical component is the visual tokenizer (VT) that maps continuous pixel inputs to discrete token sequences. The quality of the VT largely defines the upper bound of AR model performance. However, current discrete VTs fall significantly behind continuous variational autoencoders (VAEs), leading to degraded image reconstructions and poor preservation of details and text. Existing benchmarks focus on end-to-end generation quality, without isolating VT performance. To address this gap, we introduce VTBench, a comprehensive benchmark that systematically evaluates VTs across three core tasks: Image Reconstruction, Detail Preservation, and Text Preservation, and covers a diverse range of evaluation scenarios. We systematically assess state-of-the-art VTs using a set of metrics to evaluate the quality of reconstructed images. Our findings reveal that continuous VAEs produce superior visual representations compared to discrete VTs, particularly in retaining spatial structure and semantic detail. In contrast, the degraded representations produced by discrete VTs often lead to distorted reconstructions, loss of fine-grained textures, and failures in preserving text and object integrity. Furthermore, we conduct experiments on GPT-4o image generation and discuss its potential AR nature, offering new insights into the role of visual tokenization. We release our benchmark and codebase publicly to support further research and call on the community to develop strong, general-purpose open-source VTs.