Grok 4.8 Gets a ‘C++ Complex’ as Musk Tips His Hat to Claude Opus 5.5

xAI founder Elon Musk said Anthropic’s Claude Opus 5.5 is “not as good” for Grok 4.8 to match, calling the rival model “a masterpiece” in a Tuesday reply on X to a question from user Dimdv. Musk added that Grok 5 “should exceed Opus 5.5,” potentially in two or three months, but he did not provide a firm release date or further details about that prediction.

Musk said Grok 4.8 had completed pretraining and was undergoing reinforcement learning, or RL, in which model responses are refined using feedback and rewards. The team also spent considerable time cleaning up supervised fine-tuning data, or examples used to teach the model how to follow instructions. He said xAI abandoned JAX and SGLang in favor of an inference and reinforcement-learning system rewritten in C and C++.

Developer tetsuo celebrated the announcement with “C wins again,” prompting Musk to clarify that the code was “C with a little C++.” He credited Starlink engineers with work that helped produce the Grok training and inference stack, including SpaceX’s practice of writing its own headers and handling precise time calculations with leap seconds.

Musk described Grok 4.8 as “not yet incredible” but said its foundations were finally built correctly. He expects a soft rollout to Bot users later this week, ahead of a broader release. Anthropic released Claude Fable 5.2 and Mythos 5.2 on Monday, while Claude Opus 5.5 arrived on September 22 with performance comparable to the higher-end Fable 5.1 model on most work.