update on the same bundle. a new prefill kernel landed in the prismml fork this morning,…
This is a dev post classified by Jev as AI dev tools (a launch), kept by the Dev Radar because it carries real work, not commentary.
update on the same bundle. a new prefill kernel landed in the prismml fork this morning, i rebased onto it and rebuilt, and prompt processing went from 267 to 522 tok/s on the 3060. this one is purely how fast it reads your prompt. a 32k prompt that took two minutes to load now takes one. same weights, same gguf, nothing to re-download except the 492 mb bundle. i ran both builds back to back in one session rather than comparing against yesterday's number.
Posted by Sudo su (36.4k followers) 58 min ago · 4 likes · 1.1k views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- when you’re an AI maxi and GPT-6 SOL + Opus 5.5 launch on the same day — @kloss_xyz
- Impressive paper showing how much the harness changes a coding agent's results. — @omarsar0
- Opus 5.5 holy shit this is pure js code for everything music , art 🤯 — @chetaslua
- Let's delve in. 🧵 — @MarcJBrooker
- .@AnthropicAI just shipped Claude Opus 5.5! — @merge_api
- DocJev is the fastest way to classify and split complex document packets ⚡️. (and the… — @jerryjliu0
- قصة من DigitalOcean (سبتمبر ٢٠٢٦) — Managed Agents دخلت public preview: — @hazemomier
- Use GPT-6 Sol and Luna with AI SDK — @aisdk
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 21.4k posts from 5k X accounts over the last 21 days, 2.4k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 19:56 UTC. Full method.