阿猫的博客

阿猫的博客

LLM 选型

249
2024-04-02

Leaderboard

Streamlit
LLM Leaderboard 2024
LLM Rankings | OpenRouter

模型概览

model cost(input) cost(output) context
gpt-4-turbo $10/1M tokens $10/1M tokens 128k
gpt-4 $30/1M tokens $60/1M tokens 8k
gpt-4-32k $60/1M tokens $120/1M tokens 32k
gpt-3.5-turbo $0.50/1M tokens $1.50/1M tokens 16k
claude-3-haiku $0.25/1M tokens $1.25/1M tokens 200k
claude-3-sonnet $3/1M tokens $15/1M tokens 200k
claude-3-opus $15/1M tokens $75/1M tokens 200k
claude-2.1 $8/1M tokens $24/1M tokens 200k
claude-2.0 $8/1M tokens $24/1M tokens 100k
claude-instant $0.8/1M tokens $2.40/1M tokens 100k
claude-aws same same same
llama2-13b-aws $0.75/1M tokens $1/1M tokens ?
llama2-70b-aws $1.95/1M tokens $2.56/1M tokens 32k
mistral-7b-aws $0.15/1M tokens $0.2/1M tokens 8k
gemma-7b - - 8k
llama2-7b - - 4k
llama2-70b - - 32k
mistral - - 8k
grok-1(314b) - - 8k

模型能力对比

Claude vs GPT

image.png

中文能力对比

截至 2024 年 2 月。

WechatIMG37.jpg