Leo Linsky(@leo_linsky):We ran Grok 4.7 through 100 multi-agent coding evaluations. This wasn't the result I expected. When it managed to get a working build, its solutions were often more insightful than comparable frontier models, but it wasted turns and sometimes whole submissions with syntax and runtime errors, and its typical response time was much slower than Grok 4.6. Overall it's a flop.

Leo Linsky

@leo_linsky

RL and RSI @ Also: computational chemistry, quant trading, EE Generally curious

Joined March 2014

189 Following 1.1K Followers

Leo Linsky@leo_linsky

2026.09.23 03:45

We ran Grok 4.7 through 100 multi-agent coding evaluations. This wasn't the result I expected. When it managed to get a working build, its solutions were often more insightful than comparable frontier models, but it wasted turns and sometimes whole submissions with syntax and runtime errors, and its typical response time was much slower than Grok 4.6. Overall it's a flop.