aiaimodel.cards
← All benchmarks

MATH (CoT)

We haven't written a description for this benchmark yet; see the card notes below for how each maker ran it.

Your workloadRAG over docs · 12K in · 600 out · 50K requests/mo · 30% cached
  1. 1Llama 3.3 70B · Metabest value
    77.0%
  2. 2Llama 3.1 8B · Metabest value
    51.9%

“Best value”: no other model here reports a higher score at a lower cost for your workload. Scores are self-reported by each maker with different settings, so treat close calls as ties.

How each maker reported it

  • Llama 3.3 70B Benchmarks: Instruction tuned models

    Meta's own results with its internal evaluation library; shots and metrics per Meta's table.

  • Llama 3.1 8B Benchmark scores: Instruction tuned models

    Meta's own results with its internal evaluation library; shots and metrics per Meta's table.

Only each model's own maker's number is shown; scores a maker reports for rival models are left out.