DeepSeek V4 Flash 0731 Released, Google AI Leadership Shake-Up, and Other AI Model Developments
DeepSeek released its V4 Flash 0731 model, which quickly gained attention on Hacker News with 678 points and 404 comments. The model's performance on the ARC Prize benchmark was highlighted, though specific benchmark scores were not detailed in the available articles. This release underscores the ongoing competition in AI model development, with companies like ByteDance also reportedly training a mega AI model approaching Anthropic's Mythos, three times larger than Moonshot's Kimi K3.
10 stories · 6 sources
#deepseek #ai #benchmarks