AMA

Value ranking

Best value on MATH-500

500 high-school competition math problems requiring multi-step solutions. Scored on final-answer correctness.

“Value” is normalized benchmark score (0–100 for this leaderboard cohort) divided by input price per million tokens. Higher means more capability per dollar on this axis only — always sanity-check latency, context length, and your real workload.

  1. 1
    Qwen2.5 72B Instruct
    Alibaba (Qwen)
    111.11
    100.0 / $0.90/M
  2. 2
    Llama 3.3 70B Instruct
    Meta
    81.89
    72.1 / $0.88/M
  3. 3
    Llama 3.1 70B Instruct
    Meta
    53.48
    47.1 / $0.88/M
  4. 4
    Mixtral 8x22B
    Mistral
    0.00
    0.0 / $1.20/M
Open full leaderboard

AI Model Analyzer does not recommend specific vendors; rankings are derived from public data only.