Ajax: PewDiePie's fine-tuned Qwen 3.5 9B local agent
Felix Kjellberg (PewDiePie) unveiled Ajax, a Qwen 3.5 9B fine-tune powering his self-hosted Odysseus workspace for web search, browsing, email, and calendar on local hardware.
Why it matters
A celebrity-built, abliterated Qwen fine-tune aimed at a fully local personal agent — the most mainstream push yet for the run-your-own-assistant movement.
What you could build with it
An indie hacker could study Ajax's approach — distilling a small 9B model on a user's own email, calendar, and browsing patterns — to ship a privacy-first personal assistant that runs entirely on a home GPU, competing with cloud assistants on data-sovereignty grounds.
Does it hold up?
Too early to judge: as of Oct 2 there are no public weights, model card, benchmarks, or confirmed license — Ajax itself cannot be tested by third parties yet. The underlying Odysseus workspace is MIT-licensed and popular.
Built with Ajax
- PewDiePie built his own uncensored AI model. Here's what Ajax actually issteemit · Third-party analysis of Ajax's training approach, the Heretic abliteration step, and its goals as a local-first assistant.
- Anyone using Odysseus Harness? (Pewdiepie's Harness)vibingtalk.com forum · Community thread from Oct 2 of users trying the Odysseus workspace and discussing Ajax's RL-trained behavior.
Learn more
First spotted on article: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.