Jack Paul 
Elon Musk gave Anthropic’s Claude Opus 5.5 a rare endorsement on X this week while setting low expectations for his own next model. Asked how xAI’s Grok 4.8 would compare, Musk said it would not be as good. He also called Opus 5.5 a masterpiece. The comments matter because xAI is still trying to close the gap with Anthropic and OpenAI.
Grok 4.8 still in training
Musk said Grok 4.8 had finished pretraining but was still in reinforcement learning. That phase uses feedback and rewards to refine responses. He said the team also spent a lot of time cleaning supervised fine-tuning data, the examples used to teach a model to follow instructions. The update is expected to reach some users this week. Musk described it as not yet incredible, but said its foundations were finally built correctly. He plans a soft rollout to Bot users later this week before any wider release.
Code changes and Starlink help
The xAI team dropped JAX and SGLang, according to Musk, and moved to an inference and reinforcement-learning system rewritten in C and C++. Inference is how a model generates answers after a prompt. Rewriting that software can give engineers more control over performance. Musk did not share numbers showing gains. A developer celebrated the C angle. Musk clarified the code was mostly C with some C++. He also credited Starlink engineers with work that helped the Grok training and inference stack, pointing to SpaceX’s practice of writing its own headers and handling precise time calculations, including leap seconds.
Anthropic’s lead on benchmarks
Anthropic released Opus 5.5 on September 22. It performs comparably to Fable 5.1 on most work while costing 40% less to run than the previous Opus version on typical workloads, the company said. The model targets difficult programming jobs and multistep agent tasks with limited human help. On Terminal-Bench 4.0, Anthropic reported 66.4% for Opus 5.5 and 57.9% for OpenAI’s GPT-6 Astra. Grok remains far behind on public leaderboards, and Musk has not published Grok 4.8 scores that would show the size of the gap.
Grok 5 is next, maybe
Musk said Grok 5 should exceed Opus 5.5, perhaps in two or three months. He has a record of missing his own target dates, so that timeline is far from firm. For now, the near-term test is the limited Grok 4.8 rollout. The bigger challenge is still ahead.