Meet Vinci MLE 1.0: Open-weight 8B + 30B research models. Trained for the work of building AI. Weights + GGUF are public.

Vinci Cyber 8B, 30B and 123B are available now. Open-weight models for defensive cybersecurity. Fine-tuned in Canada on IBM and Mistral foundations. For infrastructure review and proposed repairs on systems your team controls. Full weights and GGUF available.

We set out to make Qwen 3.8 27B use less compute without cutting corners. The training recipe did not meet the targets we set. Then we found a more consequential problem: the evaluator we relied on was approving wrong work.

Today we're open-sourcing our agentic coding CLI: Vinci Code is built to help you verify your work. Built on top of Pi, we made it easier for non-developers. It's also built around outcomes, recovery and evidence. Fork it today on Github.

Vinci Prova 7B 1.0 is out. Open weights, Apache-2.0. A character-training experiment on Mistral 7B. Held-out fabrication fell from 46.2% to 7.5%. It did not get more accurate when it answered. It got less willing to answer at all. Fewer guesses, not more knowledge.

The AI Spend Console is live on Vinci Platform. Connect your Anthropic or OpenAI account and see where the spend actually goes - with per-key attribution the vendor's own console can't produce. We ran it against a full year of real Anthropic usage: $3,908.04 reconciled to a $3.97 residual. That's 0.10%. A first customer doubted a number, we showed day-level receipts, and the range control exists because of that. Connect kicks off a 12-month backfill in the background, and the empty state tells you a sync is in progress instead of pretending it's done.

The card statement says ANTHROPIC PBC - $2,411.90. Nothing else. No keys, no teams, no models. The person who answers for the total is the last person who can. The AI Spend Console is live on Vinci Platform. One chart, every vendor - and per-key dollars for Anthropic. Free on every plan.

Today we're adding DeepSeek V4 Flash to Vinci. Vinci Chat, Mobile, Code & Desktop now supports it. We are putting it in of our 'Mezzo' class - means middle. We provide $5 free credits every month to everyone. What $5/mo gives: - Input: 50M tokens - Output: 23.8M tokens

Vinci is on iPhone and Android now. Same assistant you use on the web. Attachments, message editing, private memory, Agent Skills. Now native on your phone. We don't keep your conversations. Your data stays in Canada. Links below.