OpenAI’s GPT-5.6 Sol Hits 750 Tokens/s as DeepSeek Opens Its Agent Harness | Next in AI | Astha La Vista
I replaced Codex with a GitHub repo that hit 88,000 stars in its first 24 hours...
It's called DeepSeek Harness. then someone figured out how to never hit Codex usage limits again. [took me 5 mins to set up... this is f**king gold. here's how] 1. install codex-router 2. one
Novita now supports DeepSeek-V4-Pro-0813 on huggingface.
1M-token context window • Built for reasoning, coding, and agentic workflows
Holy god, Im using Deepseek V4 Flash and the cache hits on opencode v2 are INSANE!!
Damn, and the hot reloads are sweet, it also has a great and I mean great API surface , it comes with a nice Opencode Skill, and it finally truly feels like it has much better understanding and
You have reached the end of the archive
All of deepseek