{"benchmark_id":"gaia2","benchmark_name":"GAIA2","benchmark_description":"GAIA2 evaluates general-purpose AI agents on real-world, multi-step questions that require reasoning, tool use, and information retrieval.","max_score":1.0,"categories":["reasoning","general","agents","tool_calling"],"modality":"multimodal","total_models":1,"entries":[{"rank":1,"model_id":"muse-glimmer-30b","model_name":"Muse Glimmer-30B","organization_name":"Meta","organization_id":"meta","benchmark_score":0.433,"normalized_score":0.433,"verified":false,"self_reported":true,"provider_id":null,"input_cost_per_million":null,"output_cost_per_million":null,"speed_rps":null,"context_window":null,"release_date":"2026-08-10","announcement_date":"2026-08-10","multimodal":true,"param_count":29600000000,"is_new":false}]}