Argon still trails the OpenAI-Anthropic duopoly
A commentator argues Gemini 4 Argon is strong but still a tier below ChatGPT and Claude, and mentions the rumored next-gen Carbon.
We really are stuck in a duopoly between OpenAI and Anthropic. Sure, Argon is a great model, but it's still not even in the same category as ChatGPT and Claude. Grok Bot is a neat UI but it feels very 6-12 months behind SOTA. Muse is bench-maxed, and I suspect Grok is too. It really seems like the competitive landscape has broken down into the two flagship true frontier companies and then everyone else who's just either distilling off them and/or focusing on post-training smaller models. However, with the leap that Google made from Pro 3.1 to Argon, perhaps their next turn of the crank will be closer to parity. The rumor mill suggests they cracked RSI too, and so Carbon is allegedly not far behind. At the same time, OpenAI and Anthropic are locked in a red queen style race, pacing each other. Their cadence has really increased. This leads me to understand why some of them have thought "whoever is first across the finish line wins forever" because it does seem like even the well-funded late-starters are having trouble catching up (Google, Meta, and SpaceX all struggling to truly reach parity). But at the same time, we're seeing market saturation and full commoditization of intelligence is on the horizon. Which basically means, within a few years, "anyone with GPUs will have frontier intelligence." At least that's my personal take.