Today, in response to rumors that ByteDance might initiate a price war for large models, Douyin Vice President Li Liang issued a statement on social media, clearly stating that this is not a price war. Li Liang pointed out that the Doubao large model has reduced costs through technological innovation, with significant optimizations in algorithms, software engineering, and hardware solutions. He mentioned that the pricing of 0.3 yuan per 1,000 tokens not only has a considerable gross profit but also follows a transparent pricing strategy, which is not the traditional 'list price discount' model.

Google announced a price cut for its fast AI model, Gemini 1.5 Flash, reducing input token costs to $0.075 per million tokens and output token costs to $0.30 per million tokens, representing a decrease of up to 78%. This move is highly beneficial for developers, particularly for scenarios such as summarization, classification, and multimodal understanding. The new API and AI studio from Google have enhanced PDF analysis capabilities, supporting the understanding of PDFs that include graphical and image content, and expanded language support to cover over 100 languages.