DeepSeek v4.1 Flash might actually be good enough for the majority of my work.
I’m spending $300/month across two Codex accounts and still constantly worrying about limits. Meanwhile, I’ve had 5+ agents running on DeepSeek all day and I’ve used 15% of my weekly allowance.
Here's Qwen 3.8 27b in action using the DeepSeek harness.
This is self-hosted from our facility. 55-65 tokens per second. Text go woosh. This is FP16 precision 262k context vision enabled. We will be adding more open weight models soon, next up will be Qwen 3.8 Flash Next
OpenAI charges ~$50 per million tokens.
DeepSeek can do it for as low as 60 cents. At the All In Summit, friedberg asks $MSFT CEO satyanadella directly: do frontier labs have the wrong business model?
Today's update: 1. Guardrails are live.
Set policies that check every request and response, and block or redact what you don't want through. 2. Server tools are live: minirouter:web_fetch and minirouter:datetime. 3. Send response_format with a JSON
You have reached the end of the archive
All of deepseek