Musk Says Grok 4.8 Trails Claude Opus 5.5
Elon Musk said xAI's upcoming Grok 4.8 model will not match Anthropic's Claude Opus 5.5, while predicting Grok 5 could surpass it.
Updated on Nov. 10, 2026 in Artificial Intelligence
Elon Musk posted on X that Grok 4.8 would not be as good as Claude Opus 5.5, calling Opus 5.5 a masterpiece. Grok 4.8 has finished pretraining and is still undergoing reinforcement learning and supervised fine-tuning data cleanup.
Why it matters
The exchange offers a public benchmark-style comparison between xAI's next model and a leading competitor's release. Musk's expectation that Grok 5 could surpass Opus 5.5 signals where xAI is directing its near-term development.
Terminal-Bench 4.0 scores measure the percentage of complicated command-line tasks a model successfully completes, with Opus 5.5 at 66.4% and GPT-6 Astra at 57.9%. Grok 4.8's benchmark performance has not been published, and Musk gave no firm release date for Grok 5.
The players
Elon Musk
Elon Musk is the billionaire entrepreneur who founded xAI and leads companies including Tesla and SpaceX, and he frequently comments publicly on AI models through his X account.
xAI
xAI is the artificial intelligence company Elon Musk founded, which develops the Grok family of large language models and integrates them into the X platform.
Anthropic
Anthropic is an AI safety company that develops the Claude family of models, including Opus 5.5, and competes directly with xAI in the frontier model market.
The details
Musk responded on X to a user named Dimdv, saying Grok 4.8 would fall short of Opus 5.5. He said the model is taking longer because the team spent considerable time cleaning up supervised fine-tuning data, and that Starlink engineers contributed to the Grok training and inference stack.
Timeline
Anthropic released Claude Opus 5.5 on September 22, 2026.
Musk posted the comparison reply on X on November 10, 2026.
Grok 4.8 is expected to have a soft rollout to Bot users later this week.
Musk estimated Grok 5 could exceed Opus 5.5 in two or three months.
The Tech Race
The Terminal-Bench 4.0 benchmark, which scores models on complicated command-line tasks, has become a reference point for comparing frontier AI systems. Musk's remarks place Grok against Anthropic's Opus 5.5 on that same measure as the frontier race intensifies.
Opus 5.5 costs 40% less to run than the previous Opus version on typical workloads, which may lower prices for developers building on the model. Users of Grok will see whether Grok 4.8 closes the gap when its soft rollout reaches Bot users.
The takeaway
Public comparisons between frontier models are increasingly tied to specific agentic benchmarks, so readers should check named scores like Terminal-Bench 4.0 rather than general claims of superiority. Musk's timelines for Grok 5 remain unconfirmed and should be weighed accordingly.
What happens next
Grok 4.8 is expected to have a soft rollout to Bot users later this week, and Musk has estimated Grok 5 could arrive in two or three months.
Further reading
For broader context on rival frontier models, see Artificial Intelligence coverage.
