Strata update! An RTX 3060 12GB got 3× faster at processing Qwen3.8-Flash-Next prompts.
This is a AI post classified by AGK Intelligence as AI infra & evals (a launch), kept by the AI Radar because it carries real work, not commentary.
Strata update! An RTX 3060 12GB got 3× faster at processing Qwen3.8-Flash-Next prompts. Strata 0.2.12 dropped, and the these speedups are worth the upgrade. 👇 🎮 RTX 3060 12GB 🧠 Qwen3.8-Flash-Next 💾 UD-IQ4_XS-v31 2K prompt processing: 🐌 Strata 0.2.11 → 87 tps 🚀 Strata 0.2.12 → 264 tps 32K prompt processing: 🐌 361 tps 🚀 981 tps And decode improved too: 💻 Coding → +37% 🤖 Agent tasks → +34% Strata discovered that its startup memory check could incorrectly count memory already used by the engine as free. So the older version sometimes moved expert reads onto a slower file-cache path unnece
Posted by David Hendrickson (12.9k followers) 54 min ago · 4 likes · 429 views · view the original post on X. Kept by the AI Radar as AI infra & evals.
More AI work like this
- LangChain Community Spotlight: Which User Is Costing You Money? 💸 — @LangChain_OSS
- Cerebras released a video of Kimi K3.1 running at 2,000+ tokens per second on its… — @rohanpaul_ai
- Running Grok Bot is living in the future. — @BrianRoemmele
- Running Grok Bot is living in the future. — @BrianRoemmele
- DeepSeek V4.1 Flash on OpenCode Go has high speed variance — @AdamHoltererer
- Free DeepSeek V4.1 Flash + GLM 5.3 Flash. — @0x0SojalSec
- The first battle of my AI tournament is here! ⚔️ — @niklaslenz_ai
- Starting work on a eval suite builder in @VulcanBench, based on my own experience that… — @morganlinton
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 65.2k posts from 5k X accounts over the last 14 days, 10.6k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-07 20:50 UTC. Full method.