Chinese AI
Tencent Hy4 is open-weight. Its 49B-active label still needs 214 GiB
Tencent Hy4 activates 49B parameters, but its smallest official low-bit build is 213.66 GiB and needs patched inference software.
Read more
Topic
Chinese AI
DeepSeek, Kimi, Qwen, and Chinese models reshaping the race.
10 articles
Musthave.AI — AI models, tools, and practical news for builders
Chinese AI
Qwen3.8-Flash-Next has 6B active parameters. It is not a 6B deployment
Qwen3.8-Flash-Next activates about 6B parameters per token but carries far more weights. Here is how to test memory, context, speed, cost and licensing.
Read more
Chinese AI
DeepSeek V4 Pro 0813 is live. August 16 rewrites the price sheet
DeepSeek V4 Pro 0813 is live across app, web, and API. New peak and off-peak rates begin August 16 and change the routing math.
Read more
Chinese AI
DeepSeek V4 Flash’s cheapest price is a cache hit. Budget the miss
DeepSeek V4 Flash keeps its low-cost agent route, but peak and off-peak rates begin August 16. Here is the new request-level math.
Read more
Chinese AI
Alibaba’s Qwen keeps climbing: Chinese models are no longer “the cheap alternative”
Chinese open models now hold most of the top open-weight slots and match the closed frontier on some coding tests. Here's how to evaluate Qwen, Kimi, and DeepSeek without the fan war.
Read more
More stories
GitHub put Kimi K3 in Copilot, then paused the rollout
Best Chinese AI models in 2026: a simple buyer’s map (Kimi, DeepSeek, Qwen, GLM)
US companies are quietly using Chinese open models — what it means for you
DeepSeek to Kimi: the Chinese open-model wave that Silicon Valley can’t ignore
Qwen inside shopping apps: when Chinese AI becomes a daily habit