in the loop

DeepSeek Coder 6.7B
#269 of 274 scored.

DeepSeekNov 2, 2023Open weights

Capabilities Index

Range 73.0–97.8

89.6

Context

Tokens

—

Input

Per 1M tokens

—

Output

Per 1M tokens

—

Benchmarks

No independent scores yet.

Epoch AI hasn't run DeepSeek Coder 6.7B yet. Its scores show up here the morning after they do.

Everything else Epoch has for it

  • GSM8K21.3%
  • Winogrande15.2%
  • MMLU15.2%
  • ARC AI215.2%

On the index

Around it.

  1. #266Baichuan1-7BBaichuan90.5
  2. #267RedPajama-INCITE-7B-BaseTogether90.4
  3. #268Dolly 2.0-12bDatabricks89.7
  4. #269DeepSeek Coder 6.7BDeepSeek89.6
  5. #270StarCoder 2 3BHugging Face88.8
  6. #271Qwen2.5-Coder-0.5BAlibaba88.2
  7. #272Cerebras-GPT-13BCerebras83.2

Sources