RLTasksby ulam.ai
Models / Qwen

Qwen / MODEL EVIDENCE

Qwen2-72B

7 benchmarks · 8 observations · 7 sources · released 2024-06-07

Scores at a glance

7 benchmarks · large dot: Qwen2-72B; small dots: closest models where reported

GPQA Diamond ↗40.78%
9.2896.4

Qwen2-72B · GPQA diamond · Best score (across scorers) · % ↗

MATH Level 5 ↗39.07%
3.2998.13

Qwen2-72B · MATH level 5 · Best score (across scorers) · % ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

4.8 percentile-point gap

5 of 6 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. Holistic Evaluation of Language Models (HELM) ↗

    Stanford Center for Research on Foundation Models

  2. GPQA Diamond ↗

    Epoch AI

  3. MATH Level 5 ↗

    Epoch AI

  4. The Agent Company ↗

    Epoch AI

  5. AI benchmarks & capabilities ↗

    Epoch AI

  6. WeirdML ↗

    Håvard Tveit Ihle

  7. Measuring AI Ability to Complete Long Software Tasks ↗

    METR

Names used in source tables

Qwen2-72B