asia-fertility 0.2.2

Tokenizer fertility, cost, and multi-turn context-budget analyzer for low-resource Asian languages.

asia-fertility 🌏

The hidden multilingual tax in your tokenizer — measured before you deploy.

asia-fertility measures the structural cost penalty that LLM tokenizers impose on lower-resource Asian languages. The same content can cost up to 11× more tokens in Burmese than in English on a frontier tokenizer — silent inflation of API bills, smaller usable context windows, and fewer in-context examples.

Quickstart

pip install "asia-fertility[oai]"

# Measure your own text
asia-fertility measure --text "தமிழ் ஒரு செம்மொழி" --lang tam --tokenizer openai/o200k_base

# Compare cost across providers (in your local currency)
asia-fertility cost --text "Xin chào" --lang vie \
  --models openai/gpt-4o,openai/gpt-3.5-turbo \
  --currencies USD,VND

# Reproduce the full 16-language × 9-tokenizer leaderboard
asia-fertility run --config configs/study_main.yaml
asia-fertility figures --run runs/main --out runs/main/figures
asia-fertility leaderboard --run runs/main --out runs/main/leaderboard.json

What's inside

• 16 lower-resource Asian languages: Vietnamese, Indonesian, Malay, Filipino, Thai, Hindi, Bengali, Sinhala, Tamil, Telugu, Kannada, Malayalam, Burmese, Khmer, Lao, plus English baseline.
• 9 tokenizers measured: OpenAI o200k_base/cl100k_base/o200k_harmony, Mistral Tekken, Qwen3, DeepSeek v3, BLOOM, Gemma-2, Aya Expanse. 3 more registered behind license walls (Llama-3.1, etc.).
• 5 metrics with 95% bootstrap CIs: fertility, premium, same-content cost ratio, characters/token (CPT), and bytes/token (BPT) — the only cross-script-fair comparator.
• NIAH benchmark: script-native needle-in-haystack across gpt-4o-mini, gpt-3.5-turbo, llama-3.1-8b-instruct on Tamil/Hindi/Burmese/Lao haystacks.

Key findings (v0.2.0)

• Same content costs 7–12× more tokens on cl100k_base for Brahmic-derived scripts (Tamil 7.61×, Burmese 11.66×).
• Switching to o200k_base cuts the penalty 3–6× (Tamil → 1.98×, Burmese → 3.18×).
• Gemma-2 is the best open-weight tokenizer for South Asian workloads (Tamil 2.58×, Burmese 4.80×).
• NIAH recall collapses to 0–7% on Hindi/Tamil/Burmese/Lao with script-native markers, even at 4k context — across all three frontier models tested. See paper §4.4.

Paper

The full writeup is at paper/paper.pdf (11 pages). Cite as:

@misc{pedretti2026asianlanguagetax,
  title  = {The Asian Language Tax: Quantifying the Cost, Context, and Recall Penalty of Tokenizing Lower-Resource Asian Languages in Frontier LLMs},
  author = {Pedretti, Antoine},
  year   = {2026},
  url    = {https://github.com/Helmo21/asia-fertility},
}

Data

The full results leaderboard + NIAH benchmark are published as a HuggingFace dataset:

from datasets import load_dataset

ds   = load_dataset("Helmo21/asia-fertility", "leaderboard")  # 144 (lang × tokenizer) rows
niah = load_dataset("Helmo21/asia-fertility", "niah")          # 536 NIAH cells

License

MIT © 2026 Antoine Pedretti. Bundled FLORES-200 data: CC-BY-SA 4.0 (Meta NLLB).