Join this channel to get access to perks:
https://www.youtube.com/@AICodeKing/join
In this video, I’ll be telling you about the new Kimi K1.5 model that is a new and fully free AI model that claims to beat Deepseek R1. Today, I’ll be testing it and we’ll see how well it performs.
—
Key Takeaways:
🔹 Kimi K1.5 is a new multimodal model available for free on its platform.
🔹 It supports both short and long chain-of-thought reasoning across multiple modalities.
🔹 The model outperforms competitors like GPT-4o and Claude Sonnet 3.5 on several benchmarks.
🔹 It uses reinforcement learning with long context scaling up to 128k for improved training.
🔹 There are two modes: a base model and a Long Thinking mode with added reasoning.
🔹 The coding performance is not as strong as Deepseek R1, but it handles general tasks well.
🔹 Although not open source, the model offers a range of features and insightful research findings.
—–
Timestamps:
00:00 – Introduction
02:49 – Testing
07:27 – Final Charts, Thoughts and More
source
