Breeze-7B
MediaTek's open 7B model optimized for Traditional Chinese and English.
About
Released March 7, 2024, MediaTek Research Breeze-7B (MR Breeze-7B) is a 7-billion-parameter open-source model built on Mistral-7B-v0.1 and continually trained for Traditional Chinese and English. Its vocabulary was expanded from 32k to 62k tokens for Traditional Chinese, with an 8k context window; MediaTek reported it halves the time and memory needed for complex Traditional Chinese inference versus comparable models.
Limitations: niche language focus means weaker performance outside Traditional Chinese and English; speed claims are vendor-reported.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 6/10
Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 67; Trending score 62/100 from researcher assessment.