Since i don't have access to Qwen4, decided to burn through my DeepSeek credits !!!
Starting with what community is showcasing for Qwen4, goal is to see how well DeepSeek-v4.1-flash result is. I tried aijoey prompt for building a real-time gravitational N-body simulation with
Same AI brain. Different coding agents.
Wildly different results. 🤯 Claude Code vs Codex vs DeepSeek Harness. A new benchmark reveals a surprising twist: there's no clear universal winner. I broke down the engineering behind it. 🎥
Was going to post this yesterday but Qwen 4 had me switch everything up.
Was really happy with this, this is Qwen 3.8 27b through Deepseek Harness all local. Took a bit but I like how it all turned out.
DeepSeek V4.1 Flash 1017 Ultrafast is now available on AI Gateway.
It generates over 1,000 tokens per second. Use deepseek/deepseek-v4.1-flash-1017-ultrafast for coding, reasoning, and agents.
You have reached the end of the archive
All of deepseek