Gemini 4 Argon
In this video, I look at the pre-announced Gemini 4 Argon. Well, I can't show outputs of this model currently because it's not officially released to the public yet. We can certainly see some of the interesting changes that Google's made to get back into the top 3 labs for frontier-level intelligence. Blog: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/ AA: https://artificialanalysis.ai/models/gemini-4-argon Sam's 2nd Channel: https://www.youtube.com/channel/UCUF7WtR4JpVbQqcQGrLeADw Twitter: https://x.com/Sam_Witteveen 🕵️ Interested in building LLM Agents? Fill out the form below Building LLM Agents Form: https://drp.li/dIMes 👨💻Github: https://github.com/samwit/llm-tutorials ⏱️Time Stamps: 0:00 Gemini 4 Argon announced 0:25 1M tokens in a single response 1:16 What Argon is built for 1:39 Artificial Analysis Intelligence Index score 2:37 1M token output vs 64K and 128K caps 3:32 Longer thinking 4:44 Generation speed 5:19 Simpler agent harnesses 7:21 Cost per task vs GPT-6 Astra 8:36 Agentic benchmarks