MIRAGE: A Benchmark for Multimodal Information-Seeking and Reasoning in Agricultural Expert-Guided Conversations

TL;DR

MIRAGE is a comprehensive benchmark designed to evaluate multimodal reasoning, decision-making, and interaction strategies in agricultural expert-guided conversations, emphasizing real-world complexity and diversity.

Contribution

It introduces MIRAGE, a large-scale, real-world, multimodal benchmark for agricultural expert interactions, covering diverse scenarios, entities, and open-world reasoning challenges.

Findings

MIRAGE includes over 35,000 interactions and 7,000 biological entities.

The benchmark challenges models with real-world, underspecified, and open-ended scenarios.

It enables evaluation of grounded reasoning, clarification, and long-form generation in agriculture.

Abstract

We introduce MIRAGE, a new benchmark for multimodal expert-level reasoning and decision-making in consultative interaction settings. Designed for the agriculture domain, MIRAGE captures the full complexity of expert consultations by combining natural user queries, expert-authored responses, and image-based context, offering a high-fidelity benchmark for evaluating models on grounded reasoning, clarification strategies, and long-form generation in a real-world, knowledge-intensive domain. Grounded in over 35,000 real user-expert interactions and curated through a carefully designed multi-step pipeline, MIRAGE spans diverse crop health, pest diagnosis, and crop management scenarios. The benchmark includes more than 7,000 unique biological entities, covering plant species, pests, and diseases, making it one of the most taxonomically diverse benchmarks available for vision-language models, grounded in the real world. Unlike existing benchmarks that rely on well-specified user inputs and closed-set taxonomies, MIRAGE features underspecified, context-rich scenarios with open-world settings, requiring models to infer latent knowledge gaps, handle rare entities, and either proactively guide the interaction or respond. Project Page: https://mirage-benchmark.github.io