Eubiota/eubiota-planner-8b

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jan 7, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

Eubiota/eubiota-planner-8b is an 8-billion parameter causal language model developed by Stanford University, fine-tuned from Qwen3-8B. It specializes in autonomous microbiome discovery, serving as the core planner for the Eubiota agentic AI framework. This model excels at orchestrating multi-agent reasoning and tool-grounded scientific exploration, particularly for microbiome hypothesis generation and experimental design.

Loading preview...

Eubiota-Planner-8B: Specialized AI for Microbiome Discovery

Eubiota-Planner-8B is an 8-billion parameter causal language model, fine-tuned from Qwen3-8B by Stanford University. It is specifically designed as the central planning module for the Eubiota agentic AI framework, focusing on autonomous microbiome discovery.

Key Capabilities

  • Agentic Planning: Orchestrates multi-agent reasoning and tool-grounded scientific exploration.
  • Microbiome Research: Generates hypotheses about host-microbe interactions and designs experiments.
  • Multi-Tool Coordination: Selects and sequences 18 domain-specific tools, including PubMed, KEGG databases, and MDIPID databases, for complex scientific queries.
  • Advanced Training: Utilizes GRPO-MAS (Group Relative Policy Optimization for Multi-Agent Systems), a reinforcement learning approach for multi-agent coordination, enabling effective tool selection and sequencing.

Training and Evaluation

The model was trained on a diverse set of domain-specific datasets, including Microbio-Bench for microbiome reasoning, PubMedQA, MedQA-USMLE for medical-biology reasoning, DeepMath-103K for mathematical reasoning, and Natural Questions for agentic search.

It is evaluated on six benchmarks, covering both general biomedical competence (MedMCQA, WMDP-Bio) and specific microbiome mechanistic reasoning tasks (Drug-Microbe Impact, Microbe-Protein Mechanism, Protein Functional Comprehension, Protein-Gene Mapping).

Intended Use Cases

  • Microbiome Hypothesis Generation: Formulating mechanistic hypotheses.
  • Experimental Design: Planning scientific workflows for microbiome research.
  • Scientific Query Resolution: Answering complex scientific questions by coordinating various domain-specific tools.