AI RADAR

AI & Vibe Coding Pulse

Daily radar of verified events: model updates, Cursor & Claude Code releases, GitHub open-source repositories, and engineering insights.

RSS feed• Andriy Orlov Telegram Channel

Thinking Machines releases Inkling 2 Small: a MoE model that competes with larger flagships - 1

Thinking Machines releases Inkling 2 Small: a MoE model that competes with larger flagships - 2

Models

1•11/06/2026, 18:31

Thinking Machines releases Inkling 2 Small: a MoE model that competes with larger flagships

1 min read

Free Course: Build a DeepSeek-style Model from Scratch

Models

150•09/27/2026, 09:32

Free Course: Build a DeepSeek-style Model from Scratch

2 min read

DeepSeek Releases DeepGEMM: Open-Source GPU Matrix Acceleration Library for AI Models

GitHub

128•10/02/2026, 10:17

DeepSeek Releases DeepGEMM: Open-Source GPU Matrix Acceleration Library for AI Models

1 min read

Open MoE Video Model MAGI-2 Preview by Sand.ai - 1

Open MoE Video Model MAGI-2 Preview by Sand.ai - 2

Models

316•08/17/2026, 19:45

Open MoE Video Model MAGI-2 Preview by Sand.ai

2 min read

Models

300•07/07/2026, 07:15

Open Compact Language Model with 50,000 Stars on GitHub

2 min read

Models

146•10/06/2026, 19:18

Mistral Unveils Le Chonk: Large 4, a Multimodal Model with 1.05 Trillion Parameters

2 min read

Release of the Day: MiMo-2.6 by Xiaomi — New Open SOTA Model - 1

Release of the Day: MiMo-2.6 by Xiaomi — New Open SOTA Model - 2

Release of the Day: MiMo-2.6 by Xiaomi — New Open SOTA Model - 3

Models

354•09/22/2026, 17:55

Release of the Day: MiMo-2.6 by Xiaomi — New Open SOTA Model

2 min read

FreeToken: A New System for Running Large MoE Models on Regular PCs - 1

Models

442•08/22/2026, 09:15

FreeToken: A New System for Running Large MoE Models on Regular PCs

3 min read

Qwen-4 Architecture: New Achievements in AI - 1

Qwen-4 Architecture: New Achievements in AI - 2

Qwen-4 Architecture: New Achievements in AI - 3

Models

307•08/26/2026, 18:15

Qwen-4 Architecture: New Achievements in AI

2 min read

DeepSeek and Huawei Release Open-Source CUDA Alternative for Ascend Chips

Models

158•09/30/2026, 17:48

DeepSeek and Huawei Release Open-Source CUDA Alternative for Ascend Chips

2 min read

Apple Releases M6 and M5 Ultra - 1

Apple Releases M6 and M5 Ultra - 2

Models

438•08/25/2026, 23:00

Apple Releases M6 and M5 Ultra

2 min read

Qwen 3.8-Max Released with 2.4 Trillion Parameters

Models

318•08/03/2026, 09:45

Qwen 3.8-Max Released with 2.4 Trillion Parameters

1 min read

Aleph Alpha Releases Kolibri-1: Open-Weights German AI Model with 78B Parameters - 1

Aleph Alpha Releases Kolibri-1: Open-Weights German AI Model with 78B Parameters - 2

Aleph Alpha Releases Kolibri-1: Open-Weights German AI Model with 78B Parameters - 3

Models

169•10/05/2026, 08:32

Aleph Alpha Releases Kolibri-1: Open-Weights German AI Model with 78B Parameters

2 min read