ctxdexSynthetic model data for learning

Benchmark family

LiveBench

A family groups the versions of one benchmark. It is not evaluated and has no applicable models of its own; each of its 12 definitions is resolved separately.

General reasoningRolling evaluation

Definitions

Compare the newest 4 →

Newest first, following the recorded successor chain. Each definition is resolved against models separately.

  1. BenchmarkLiveBench2026-06-25
    ActiveNewest in family
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  2. BenchmarkLiveBench2026-01-08
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  3. BenchmarkLiveBench2025-12-23
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  4. BenchmarkLiveBench2025-11-25
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  5. BenchmarkLiveBench2025-05-30
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  6. BenchmarkLiveBench2025-04-25
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  7. BenchmarkLiveBench2025-04-02
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  8. BenchmarkLiveBench2024-11-25
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  9. BenchmarkLiveBench2024-08-31
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  10. BenchmarkLiveBench2024-07-26
    Superseded
    Successor of LiveBench
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  11. BenchmarkLiveBench2024-06-24
    Superseded
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    30
  12. BenchmarkLiveBench
    Active
    Scale
    Not recorded
    Primary metric
    Accuracy
    Applicable models
    0

Family methodology

No family-level methodology recorded.

Across definitions

Tasks
AMPS Hard, Coding Completion, Connections, CTA, Code Generation (LCB), Math Competition, Olympiad, Paraphrase, Plot Unscrambling, Simplify, Story Generation, Summarize, Table Join, Table Reformat, Typos, Web of Lies v2, Zebra Puzzle, Spatial, Web of Lies v3, Code Completion, Code Generation, JavaScript, Python, TypeScript, Theory of Mind, Logic with Navigation, Consecutive Events, Integrals with Game
Sources
None recorded at family level
Last reviewed
2026-09-22

Applicability does not mean evaluated; benchmark result/model score data is outside this UI contract.