AI4SGI/ExoMind-9B

VISIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 13, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

AI4SGI/ExoMind-9B is a 9 billion parameter language model developed by the ExoMind Team at Shanghai Artificial Intelligence Laboratory, fine-tuned from Qwen3.5-9B. It is designed for scientific reasoning and agentic research, integrating the model with specialized interaction objects and autonomous interaction processes. This compact checkpoint supports resource-conscious experimentation in scientific question answering, mathematical reasoning, and tool-use, while retaining the multimodal capabilities of its base model.

Loading preview...

What is AI4SGI/ExoMind-9B?

AI4SGI/ExoMind-9B is a 9 billion parameter language model developed by the ExoMind Team at Shanghai Artificial Intelligence Laboratory. It is a compact checkpoint fine-tuned from Qwen3.5-9B, specifically designed for scientific reasoning and agentic research. The model implements an extended-mind-inspired approach, integrating the language model with specialized interaction objects and autonomous interaction processes to form a cohesive system.

Key Capabilities

  • Scientific Interaction: Supports workflows for source discovery, evidence grounding, executable verification, and observation integration.
  • Progressive CoI Training: Developed with intrinsic reasoning and interaction behaviors derived from selected pure-reasoning and interaction trajectories.
  • Multimodal Foundation: Retains the image-text capabilities inherited from its Qwen3.5 base model.
  • Resource-Conscious: Optimized for lower-resource experimentation in scientific domains.

Good For

  • Scientific Question Answering: Excels in answering complex scientific queries.
  • Mathematical and Computational Reasoning: Designed to handle intricate reasoning tasks.
  • Tool-Use Experiments: Facilitates research and development involving external tools.
  • Code-Assisted Verification: Supports verification processes aided by code execution.
  • Agentic Prototypes: Ideal for developing and testing resource-conscious AI agents in scientific contexts.