Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
Strata 0.2.4.1: DeepSeek setup works from the installed Terminal
Fixed: setting up DeepSeek V4.1-Flash from the installed Terminal stopped at "Preparing your model" with "can't open file .../tools/dsv41/pack_gguf.py". The Terminal and Control packages of 0.2.4 did not contain the DeepSeek preparation scripts (pack_gguf.py, smoke.py and the short tune, tune.py); they ran only from a source checkout. They are in the packages now, and so are three scripts that other commands run by path (strata_inspect.py, strata_bench.py, nvfp4_local_pack.py).
Only the Terminal's Control changed (version 0.1.49.1). The app stays 0.2.3 and every engine package stays as it was, so there is nothing new to download for them.
Corrected speed figure for the RTX 3060
The 0.2.4 notes said DSpark raised an RTX 3060 PC from 6.1-6.7 to 8.2-8.7 tokens/s. That was measured with DSpark forced on; on 12 GB cards DSpark stays off by default. Measured with the release defaults on an RTX 3060 PC with 46 GB of RAM: 8 to 10 tokens/s for answers (code 9.9, agent work 7.8, chat 8.9) and 149 tokens/s for an 8K prompt. Turning DSpark on (--dspark on) adds 2 to 13% there.
Updating
- From 0.2.4: run
strata updatein the Terminal. It fetches Control 0.1.49.1; the nextstratacommand already runs it (strata --versionsays 0.2.4.1). Then run your DeepSeek setup again. - New installs: the install command (
install.ps1on Windows,install.shon Linux) installs this version. - Mac: the Mac Terminal gets Control 0.1.49.1 too. DeepSeek is not on the Mac yet.
Strata is free and open source. If it runs well on your PC, a coffee keeps the work on it going:
