@kedr.bit: When the free preview for ox alpha ended and GLM 5.3 Flash launched at list prices of $0.15 per million input tokens and $0.50 per million output, it was still cheaper than DeepSeek V4 Flash on blended cost due to its aggressive cached-input rate-GLM's blended cost came in at roughly $0.10 per million tokens versus DeepSeek's $0.23, making GLM about 2.3x cheaper for cache-heavy workloads 1 6. That cost advantage, combined with a higher intelligence index score, made GLM 5.3 Flash the more attractive option, until DeepSeek released V4.1 Flash in September 2026 with dramatically reduced off-peak pricing and native multimodal support. This picture depicts deepseek v4 flash and ox alpha/FLASH 5.3 drawn as anime girls, where deepseek is annoyed about being replaced. #fyp #zai #deepseek #ai #glm