HJ Blog > Code Snippet #344: Ollama model list with context sizes in Fish

2026-07-07 05:00:27 Ollama model list with context sizes in Fish

Ollama doesn't list its models with context size, and I wanted to be able to see this info (and model size) all in one go. Fish shell is elegant for tasks like this.

( sh ) ✂

#!/usr/bin/env fish

echo
echo "Model Name                  Context    Size"
echo

for model in (ollama list | sort | egrep -v NAME)
    string replace -ra "\s+" ' ' $model | read -l name x size magnitude x x x
    set length (ollama show $name | awk '/context length/ { print $3 }')
    
    printf "%-26s  %7.0f   %5.1f %2s\n" $name $length $size $magnitude
end

echo

  1 #!/usr/bin/env fish
  2 
  3 echo
  4 echo "Model Name                  Context    Size"
  5 echo
  6 
  7 for model in (ollama list | sort | egrep -v NAME)
  8     string replace -ra "\s+" ' ' $model | read -l name x size magnitude x x x
  9     set length (ollama show $name | awk '/context length/ { print $3 }')
 10     
 11     printf "%-26s  %7.0f   %5.1f %2s\n" $name $length $size $magnitude
 12 end
 13 
 14 echo

Curiously, and perhaps but a tiny speed bump on the road to the future, ChatGPT, Claude, and my local models, all failed hard, even after many attempts, to know how to properly do line 8, where it takes the $model line and directly assigns its parts to various variable names.

To wit, the above code was completely written without any LLM help. I have yet to see a compelling case for LLM code, but perhaps some day.

RSS Feed for this page