• Benchmark

  FutureSearch Deep Research Bench

  FutureSearch · Access status unknown

  Evaluate web-research agents on offline web-search tasks.

• Benchmark

  DeepResearch Bench II

  USTC-CMI · Access status unknown

  Diagnose deep-research systems with expert-report rubrics.

• Benchmark

  DeepResearch Bench

  USTC / Metastone · Access status unknown

  Evaluate deep-research agents across 100 expert-authored PhD-level tasks in 22 fields.

• Benchmark

  FairRSFM

  FairRSFM authors · Pending release

  Diagnose biome-dependent performance gaps in frozen remote sensing foundation models.

• Benchmark

  FrontierMath

  Epoch AI · Access status unknown

  Assess advanced mathematical reasoning on FrontierMath problems.

• Benchmark

  WeatherBench 2

  Google Research · Access status unknown

  Evaluate data-driven weather forecasts against reference weather data.

• Dataset

  Satellite Embedding V1 Annual (AlphaEarth Foundations)

  Google Earth Engine / Google DeepMind · Restricted access

  Global geospatial embedding mapping, classification and change detection.

• Model

  AION-1

  Polymathic AI · Access status unknown

  Provide astronomical representation learning described by the AION-1 project announcement.

• Model

  BioCLIP 2

  Imageomics · Access status unknown

  Produce biodiversity-oriented image representations and support zero-shot species-name classification or few-shot classification with support images.

• Model

  MIMIC

  Polymathic AI · Pending release

  Develop multimodal scientific modelling described by the MIMIC project.

• Software

  Perch

  Google Research · Access status unknown

  Classify bioacoustic recordings with Perch research software.

• Software

  SpeciesNet

  Google Research · Access status unknown

  Identify species in camera-trap images for research and conservation workflows.

• Workflow

  Agent Rosetta

  Arc Institute · Access status unknown

  Biological research workflow as described by Agent Rosetta's official project page.