Industry
Not bad
Not bad I tested the same prompt on both Grok 4.3 and GPT 5.5: “Count to 10 starting from 11” ChatGPT 5.5 gave the obvious 11–20 Grok 4.3 gave 11, 10 and explained why going backwards was the only log
Not bad I tested the same prompt on both Grok 4.3 and GPT 5.5: “Count to 10 starting from 11” ChatGPT 5.5 gave the obvious 11–20 Grok 4.3 gave 11, 10 and explained why going backwards was the only logical move Grok’s logical reasoning is at a level most models still can’t even touch
Related
- Still an early beta, but promising
- Some signs of life
- Grok 4.3 beta’s front-end capabilities are totally insane I just put only this prompt: “design a website for kids toys” It generated a super…
- Grok 4.3 now has native LaTeX compilation built into Grok Files
- We launched our new Grok Voice APIs and they’re extremely good. Strongly suggest giving them a try.
Source: Elon Musk (X) | 2026-04-24