xAI launches its latest flagship inference model, Grok 4.3.

xAI has launched the Grok 4.3 model, positioned as a practical transitional version. The model's API price has been reduced by 40%-60%, output speed reaches 196 tokens/s, supports long contexts of up to 1 million tokens, and shows significant improvements in proxy tasks and office assistance capabilities, excelling at generating documents, spreadsheets, and presentations. However, the model lags behind GPT-5.5 and Claude Opus 4.7 in the Intelligence Index, exhibiting insufficient stability in complex reasoning and fact-checking, and a slightly increased illusion rate.