Wu Haoning

@HaoningTimothy

Husband. Curious RL researcher. Humbly contributed to Joint RL of K3 and K2.5. Prev: PhD NTU, BS PKU.

Singapore

Pinned
Wu Haoning
@HaoningTimothy
Jan 27
We are really taking a long time to prove this: everyone is building big macs but we bring you a kiwi🥝 instead. You have multimodal with K2.5 everywhere: chat with visual tools, code with vision, generate aesthetic frontend with visual refs...and most basically, it is a SUPER
Kimi.ai
@Kimi_Moonshot
Jan 27
Kimi K2.5 has arrived! 🥝 Here are 2 things to know: Aesthetic Coding x Agent Swarm.
00:00
Wu Haoning
@HaoningTimothy
17h
Nights in Beijing 呼喚城門開
Wu Haoning
@HaoningTimothy
Aug 25
paloma and korrine are two models? Just like their predecessors?
Wu Haoning
@HaoningTimothy
Aug 21
key fact: the goal of stealth models in 2026 is to maximize advertisement that plain release by the lab CANNOT achieve given this fact, even disappointing, maybe we need to accept that they usually DO NOT come from the lab or IS NOT the chosen one most people hope for
Wu Haoning
@HaoningTimothy
Aug 21
Grab to see GLM (maybe? ox-alpha) and DeepSeek have finally published their Pareto-frontier-level VLMs Hope they be OSS so we are having more 🥝 in OSS model world
DeepSeek
@deepseek_ai
Aug 21
DeepSeek-V4-Flash-Vision-Exp is now live on the DeepSeek API Platform! 🚀 🔹 This experimental multimodal model matches DeepSeek-V4-Flash on text capabilities—including agents, reasoning, and world knowledge. 🔹 On multimodal agent benchmarks, V4-Flash-Vision-Exp makes a major