[Tencent Cloud: Official Version of DeepSeek-V4 'Factory Direct Supply' Model Scheduled for Mid-July Launch] According to a Tencent Cloud announcement, the Tencent Cloud large model service platform TokenHub and the intelligent agent development platform are scheduled to officially launch the official version of the DeepSeek-V4 'Factory Direct Supply' model in mid-July, along with the introduction of a peak-valley pricing mechanism. The pricing for DeepSeek-V4-Pro during regular periods is as follows: - Cache hit: ¥0.025 per million tokens - Inference input: ¥3 per million tokens - Inference output: ¥6 per million tokens During peak periods, the pricing is: - Cache hit: ¥0.05 per million tokens - Inference input: ¥6 per million tokens - Inference output: ¥12 per million tokens For DeepSeek-V4-Flash, the regular pricing is: - Cache hit: ¥0.02 per million tokens - Inference input: ¥1 per million tokens - Inference output: ¥2 per million tokens During peak periods, the pricing is: - Cache hit: ¥0.04 per million tokens - Inference input: ¥2 per million tokens - Inference output: ¥4 per million tokens.