sumitasok

Array languages like J and concise dynamic languages like Python, Ruby, and Clojure use the least tokens for AI-assisted development. [1, 2, 3]

Top Token-Efficient Languages

• J (Array Language): Dominates benchmarks at around 70 tokens per task. It achieves extreme conciseness like APL but uses standard ASCII characters instead of rare mathematical glyphs, which keeps tokenizer overhead low.
• Clojure & Ruby: Average around 109 to 119 tokens per task. Dynamic and functional paradigms skip the verbose boilerplate and explicit type declarations required by languages like Java or C.
• Python: Averages around 128 tokens per task. It offers a strong balance of high semantic density, minimal syntax, and widespread LLM training data familiarity.
• Haskell: The most token-efficient statically typed language (~130 tokens), benefiting from advanced type inference that reduces redundant code. [1, 2, 3, 4, 5]

Least Token-Efficient Languages

sumitasok

(venv) ╭─sa001780 at L04M24KMF7 in ~/code/samples/coredns-on-kind-cluster on main✘✘✘ 26-02-20 - 21:24:47
╰─(venv) ⠠⠵ kind create cluster --config kind-config.1.33.1.yaml
Creating cluster "1331" ...
 ✓ Ensuring node image (kindest/node:v1.33.1) 🖼
 ✓ Preparing nodes 📦 📦
 ✓ Writing configuration 📜
 ✓ Starting control-plane 🕹️
 ✓ Installing CNI 🔌
 ✓ Installing StorageClass 💾
 ✓ Joining worker nodes 🚜

sumitasok

package migration

type Migrator interface {
	Migrate() error
}

sumitasok

# cd to the folder where you have your .mov video files.
find . -name \*.mov -print0 | xargs -0 -I{} ffmpeg -i {} -f mp4 -vcodec libx264 -preset fast -profile:v main -acodec aac -vf scale=1024:576 {}.mp4

sumitasok

{
  'type': 'RGB'
  'bandInfo': {
    0 : { 'type': '...', 'colorinterpreter': 'R' }
    1 : { 'type': '...', 'colorinterpreter': 'G' }
  }
  'cogInfo': {}
  'mapTile': {}
}

sumitasok

FROM python:3.6-slim-stretch

ADD requirements.txt /tmp/requirements.txt

RUN apt-get update && \
    apt-get install -y \
        build-essential \
        make \
        gcc \
        locales \

sumitasok

package models_test

import (
	"testing"

	"github.com/stretchr/testify/assert"
)

func TestMeta_ToJson(t *testing.T) {
	assert := assert.New(t)

sumitasok

# Class variables with @ are accessible from class methods!
class MyClass
  @var = 2
  
  def self.square
    @var *= @var
  end
end
 
MyClass.square

sumitasok

req = RestClient::Request.new(url: "www.google.com", method: 'get')
net_http_resp = Net::HTTPResponse.new(1.0, 200, "OK")
net_http_resp.add_field 'Set-Cookie', 'Monster'
resp = RestClient::Response.create(response_body.to_json, net_http_resp, req)

sumitasok

# * 100,000,000-line file generated by seq 100000000 > test.in
# * Reading lines 50,000,000-50,000,010
# * Tests in no particular order
# * real time as reported by bash's builtin time
# https://unix.stackexchange.com/questions/47407/cat-line-x-to-line-y-on-a-huge-file
awk 'NR >= 57890000 && NR <= 57890010' /path/to/file

awk 'NR < 57890000 { next } { print } NR == 57890010 { exit }' /path/to/file

tail -n+50000000 test.in | head -n10