{
  "name": "Core-30",
  "version": 1,
  "frozen": "2026-07-29",
  "purpose": "A capability-balanced, budget-conscious comparison set. It is not a proportional sample of the full 130-problem benchmark.",
  "scoring": {
    "judge": "cli/codex:gpt-5.5:high",
    "render_fail_score": 0,
    "primary_aggregate": "equal-weight mean of the Regular, Frontier, and Reconstruction group means"
  },
  "selection_policy": [
    "Freeze membership before generating the 11 expansion outputs.",
    "Use 10 problems from each top-level capability group.",
    "Within Regular and Frontier, span prior-model difficulty and distinct mathematical or visual capabilities.",
    "Compare models only on the same 30 problems with the same judge."
  ],
  "groups": {
    "Regular": [
      "archimedean_spiral_galaxy",
      "geometric_cube",
      "hopf_fibration_base_loops",
      "hyper_menger_cube_3sphere",
      "mandelbulb_fractal",
      "sierpinski_triangle_6_iterations",
      "reaction_diffusion_patterns",
      "penrose_tiling_p3",
      "number_theory_music",
      "regular_dodecahedron"
    ],
    "Frontier": [
      "coxeter_reflection_kaleidoscope",
      "minimal_surface_knot_boundaries",
      "optimal_transport_mass_flow_tubes",
      "reaction_diffusion_nonorientable_surfaces",
      "cellular_potts_tissue_folding",
      "navier_stokes_vortex_reconnection",
      "error_correcting_code_decoding_landscape",
      "qec_threshold_phase_diagram",
      "differentiable_rendering_ambiguity_landscape",
      "polyrhythm_phase_torus"
    ],
    "Reconstruction": [
      "reproduce_image",
      "reproduce_image_andrew_pons",
      "reproduce_image_fabrice_villard",
      "reproduce_image_jason_leung",
      "reproduce_image_javier_penas",
      "reproduce_image_mark_basarab",
      "reproduce_image_photoholgic",
      "reproduce_image_rayul",
      "reproduce_image_sebastien_gabriel",
      "reproduce_image_tim_stief"
    ]
  },
  "new_fable_generation_problems": [
    "reaction_diffusion_patterns",
    "penrose_tiling_p3",
    "number_theory_music",
    "regular_dodecahedron",
    "reaction_diffusion_nonorientable_surfaces",
    "cellular_potts_tissue_folding",
    "navier_stokes_vortex_reconnection",
    "error_correcting_code_decoding_landscape",
    "qec_threshold_phase_diagram",
    "differentiable_rendering_ambiguity_landscape",
    "polyrhythm_phase_torus"
  ]
}