Skip to content
← Models
Open weightsTaiwan NSTC/10 trustMetrics-checked

TAIDE

Taiwan's sovereign Traditional Chinese LLM — Llama-based, trained on local data.

Visit official page →

About

TAIDE (Trustworthy AI Dialogue Engine) is Taiwan's government-backed sovereign LLM, developed by the National Science and Technology Council starting April 2023 and released in April 2024. Built on Llama-2/3 (70B scale) and refined on carefully filtered Traditional Chinese data — government publications, local news, university resources — TAIDE is designed to serve Taiwan's population in local language patterns and resist the cultural and political biases of imported models, a concern Taiwanese researchers frame in national-security terms.

It is a 2024-era project at modest scale: a 70B Llama derivative can't match 2026 frontier models, and the related open Taiwan-LLM community models (MiuLab, 7B-70B, Llama-3 license) are the more practical artifacts for builders today. But as a case study in state-led AI sovereignty — $7.4M budgeted, explicitly motivated by resilience against foreign information influence — TAIDE is the reference implementation alongside France's Lucie and Saudi's ALLaM.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

April 15, 2024

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

chinese-nlpsovereign-aitaiwanresearch

Trust score — 6/10

/10 trustMetrics-checked

Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 54; Trending score 52/100 from researcher assessment.

Sources