Seed1.5-VL Technical Report

TL;DR

Seed1.5-VL is a compact yet powerful vision-language model that achieves state-of-the-art results across numerous benchmarks and excels in multimodal reasoning and agent-centric tasks, demonstrating broad applicability.

Contribution

This report introduces Seed1.5-VL, a novel multimodal model combining a 532M vision encoder and a 20B-parameter MoE language model, with comprehensive insights into its design, training, and performance.

Findings

Achieves state-of-the-art on 38 out of 60 benchmarks.

Outperforms leading multimodal systems in GUI control and gameplay.

Demonstrates strong reasoning abilities in visual puzzles.

Abstract

We present Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning. Seed1.5-VL is composed with a 532M-parameter vision encoder and a Mixture-of-Experts (MoE) LLM of 20B active parameters. Despite its relatively compact architecture, it delivers strong performance across a wide spectrum of public VLM benchmarks and internal evaluation suites, achieving the state-of-the-art performance on 38 out of 60 public benchmarks. Moreover, in agent-centric tasks such as GUI control and gameplay, Seed1.5-VL outperforms leading multimodal systems, including OpenAI CUA and Claude 3.7. Beyond visual and video understanding, it also demonstrates strong reasoning abilities, making it particularly effective for multimodal reasoning challenges such as visual puzzles. We believe these capabilities will empower broader applications across diverse tasks. In this report, we mainly provide a comprehensive review of our experiences in building Seed1.5-VL across model design, data construction, and training at various stages, hoping that this report can inspire further research. Seed1.5-VL is now accessible at https://www.volcengine.com/ (Volcano Engine Model ID: doubao-1-5-thinking-vision-pro-250428)