AI Model Rankings
Leading models available on BazaarLink, ranked by the Artificial Analysis Intelligence Index
X-axis: TWD input price per 1M tokens; the upper-left combines stronger capability with lower cost
Full leaderboard
Top 10 models callable on BazaarLink, priced in TWD with invoice options
Frequently asked questions
How is the Artificial Analysis Intelligence Index defined?
It is an independent 0–100 composite score from Artificial Analysis that combines public benchmarks including MMLU-Pro, GPQA Diamond, Humanity's Last Exam, AIME, LiveCodeBench and SciCode. Higher scores indicate stronger overall capability; this page uses its public data and updates daily.
How is output speed measured?
Output speed is the median number of tokens generated per second in Artificial Analysis testing. A higher tok/s value means faster generation and is distinct from time to first token (TTFT).
How can I use these models in Taiwan, and what do they cost?
Call them through BazaarLink's OpenAI-compatible API. Prices are displayed in TWD, and eligible business payment channels provide invoices. Use the leaderboard and model detail pages for current unit prices.
Does this ranking measure Traditional Chinese ability?
Not directly. The Intelligence Index is a mostly English global benchmark of general reasoning, mathematics and coding; it does not specifically test Traditional Chinese comprehension or Taiwan-local knowledge.
Intelligence and speed data: Artificial Analysis. TWD prices are BazaarLink's live selling prices and update daily.（Artificial Analysis・1 USD = 31.56 TWD）