🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!
🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!
Check out the configuration details in our official API docs: api-docs.deepseek.com/quick_…
Who would’ve believed we can run Opus 4.8 at home?
DeepSeek V4 Flash already hits ~110 tok/s at $0.14/$0.28 🤯
Opus 4.8 barely does ~60 tok/s at $5/$25..
and only matches Flash’s speed in Fast mode at double the price ($10/$50), not to mention the penalty of parallel agents running in this case.. and the dirt cheap cache rate on DeepSeek
American labs gotta cut the profit margins and drop the greed..
DeepSeek already runs on huwawei ascend.. USA is losing the AI accelerator race
🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!
🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!
Check out the configuration details in our official API docs: api-docs.deepseek.com/quick_…
funny how the ralph loop dropped almost a year ago
right when stupid startups raised millions for slack bots to keep agents running for hours
now both claude code and codex ship /goal like they invented it
the industry doesn't move as fast as the hype
Anthropic should already warn users about hypomanic episodes that could occur out of amazement at their models' capabilities
The combo of Claude Code + Superpowers + proper toolset genuinely freak me out.. Not sure if I'm still in control..
