AITechNodeabout 1 hour ago
DeepSeek puts V4-Flash API into public beta
1 min read
DeepSeek put its V4-Flash API into public beta, upgrading it for agent tasks. The model scored 82.7 on Terminal Bench 2.1 and 54.4 on DeepSWE. It adds Responses API support and Codex adaptation. V4-Flash-0731 keeps the preview's structure and size but was retrained. V4-Pro and app models remain unchanged.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
