The best uncensored xAI model is Grok 4.5 (53.8 intelligence) — served without a provider-level moderation layer. Grok 4.3 (37.6) and Grok 4.20 (37.0) round out the top three.
Large language models served without a provider-level moderation layer, ranked by intelligence. "Uncensored" here means the API endpoint doesn't add its own filtering on top of the model — each model's own alignment training still applies, and open-weight entries can be self-hosted for full control. Relevant for fiction, roleplay, red-teaming and safety research, where an extra moderation layer causes over-refusal on legitimate prompts.
The best uncensored xAI model is Grok 4.5 (53.8 intelligence) — served without a provider-level moderation layer. Grok 4.3 (37.6) and Grok 4.20 (37.0) round out the top three.
Grok 4.3 (37.6) is the closest alternative on this metric, followed by Grok 4.20 (37.0). See the full ranking above for the tradeoffs.
modelgrep tracks 5 xAI models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Grok 4.5. 3 of them qualify for this ranking.