in the loop

Claude 3 Haiku
#206 of 274 scored.

AnthropicMar 7, 2024Closed

Capabilities Index

Range 110.5–121.5

118.3

Context

Tokens

—

Input

Per 1M tokens

—

Output

Per 1M tokens

—

Benchmarks

How it scores.

  • GPQA Diamond15.1%

    164th of 186

  • Mock AIME1.7%

    166th of 176

Everything else Epoch has for it

  • MMLU65.1%
  • ScienceQA62.7%
  • Winogrande48.4%
  • DTBench16.9%
  • MATH level 514.9%
  • CadEval12.0%
  • LMCA10.4%
  • WeirdML9.8%

On the index

Around it.

  1. #203GPT-3.5 Turbo (Nov 2023)OpenAI118.5
  2. #204Qwen2.5-7BAlibaba118.5
  3. #205Mixtral 8x7BMistral118.5
  4. #206Claude 3 HaikuAnthropic118.3
  5. #207Ministral 3BMistral118.1
  6. #208Yi-34B01.AI117.4
  7. #209phi-3-mini 3.8BMicrosoft117.4

Sources