Grok Bot의 업무 자동화, Qwen 모델 확장, 소형 추론 모델 효율화

Min Choi

@minchoi

Grok Bot just killed 1,000 startups. Each Bot gets its own computer. It learns your workflows, uses your apps, and works while you sleep. Grok just became your 24/7 AI employee.

Yun-Ta Tsai

@yunta_tsai

Grok Bot was able to a) go through the calendars and find out anything I need to make reservations for beforehand that I haven’t done yet, b) determine the best time to make reservations, and c) navigate the reservations on a website. While I was walking in the parking lot before getting to my cars, I was talking to it in mixed Chinese and English. Color me impressed.

From #22 to #4 on Legal Research Bench! Solid progress for Qwen3.8-Max. Thanks for highlighting~

Try Qwen-Image-3.0 on @openart_ai !

Ant Group has released Ling 3.0 Tiny, which scores 25 on the Artificial Analysis Intelligence Index. With just 7.9B total and 1.3B active parameters, it sits on the Pareto frontier for Intelligence vs. Active Parameters @AntLingAGI has released Ling 3.0 Tiny, an open weights reasoning model with a 262K-token context window. It scores 25 on the Artificial Analysis Intelligence Index, comparable to gpt-oss-120b (high, 24) with 15x fewer total parameters and 4x fewer active parameters. However, this parameter efficiency comes with high token usage: Ling 3.0 Tiny used 213M output tokens to run the Intelligence Index, nearly as many as the larger Ling 3.0 Flash (240M). It follows the recent release of Ling 3.0 Flash, which scores 38 with 124B total and 5.1B active parameters. The weights are available on Hugging Face under an MIT license. Key results: ➤ Ling 3.0 Tiny extends the open weights Pareto frontier for Intelligence vs. Active Parameters. It scores 25 on the Artificial Analysis Intelligence Index, and at 7.9B total and 1.3B active parameters, the model is small enough to run locally in many settings ➤ Ling 3.0 Tiny makes a 59-point improvement in AA-Omniscience over Ling-mini-2.0, primarily driven by improvements in hallucination rate. Ling 3.0 Tiny scores -19 with 9% accuracy and a 30% hallucination rate, similar to Qwen3.6 27B (-20), but its accuracy is relatively low among similarly sized open weights models. Compared to the previous generation, where Ling-mini-2.0 had a 96% hallucination rate, Ling 3.0 Tiny improves on hallucination while maintaining the same accuracy. Instead of guessing when it doesn’t know, it attempted just 37% of questions in the AA-Omniscience evaluation. ➤ Ling 3.0 Tiny is relatively capable in tool use but performs modestly in agentic work tasks. Ling 3.0 Tiny scores 21% on 𝜏³-Banking, ahead of Qwen3.6 27B (17%). On GDPval-AA v2, however, it reaches an Elo rating of 772, behind Nemotron 3.5 Lightning at 824 despite sitting slightly ahead on the overall Intelligence Index. Additional model details: ➤ Size: 7.9B total parameters, 1.3B active (mixture of experts) ➤ Context window: 262K tokens ➤ License: MIT ➤ Providers: InclusionAI first-party API and @novita_labs

Teknium

@Teknium

Master planned

Day 251: Good Bugs First of all, check out these recovery clips. Snappy as hell. Brings a tear to my eye. It's shockingly robust. Thanks @ChenTessler for the help there First training attempt went well. But bugs. I discovered a 90 degree mismatch between the pelvis keyframe and joystick controls, so unfortunately no cool climbing clip tonight. Oh well. The @wandb logs showed some success on synthetic episodes. So with the control inputs aligned, i'm quite excited to see how the next training attempt goes. So far, my crazy idea is holding up.