Depaza Advanced
Maximum capability, for the hardest problems.
Our most powerful EU-hosted model. It reasons before it answers — and stays with a problem until it's actually solved. When getting it exactly right matters more than getting it instantly, this is the one you reach for.
Built for the problems that don't have easy answers.
Advanced is tuned for depth over speed. It works a problem through step by step, holds far more in mind at once, and is at its best exactly where a quick answer falls apart.
Deep reasoning
It thinks before it answers — planning, checking its own work and weighing options instead of blurting the first plausible reply.
Long, multi-step work
Stays coherent across long, multi-step tasks — breaking a big job into parts and holding the thread from start to finish.
Serious code
Large refactors, subtle bugs and polished front-ends — the coding work where a near-miss still costs you an afternoon.
Dense analysis
Contracts, financial statements, research and thorny logic — reasoning that has to be right, with the working shown.
Long context
Holds whole documents and codebases at once, so it reasons over the full picture instead of a keyhole view.
Correctness first
Tuned to get it exactly right rather than merely fast — the trade you want on the work that actually matters.
A frontier model is just the engine. We built the rest.
Raw model quality is table stakes. What makes Advanced genuinely useful is everything Depaza wraps around it — the search, the memory, the tools and the EU guarantees that turn a clever answer into finished work.
Live search intelligence
Our own EU-hosted search lets Advanced pull fresh facts, read the real pages and cite them — grounded in today, with no invented links or dates.
Memory & context
It remembers the files you shared and the thread you're on, and carries the important parts forward across long sessions — so you're not re-explaining yourself every few turns.
A full tool-belt
Native PowerPoint, Word and Excel, polished PDF reports, company-financials and domain lookups, image generation, vision and OCR — real deliverables, not just a wall of text.
A team behind the answer
Advanced is one specialist in a routed EU fleet. In Expert mode a second model reviews and challenges the first before you ever see the answer.
EU-hosted by default
Every request runs on EU infrastructure under EU law. Your prompts, files and code never cross the Atlantic — GDPR and EU AI Act ready.
Frontier, minus the frontier bill
Top-tier capability at a European price, with a transparent 3× token cost so you always know what a hard question is worth — and no per-seat frontier premium.
It reasons before it replies.
Most models answer the instant you hit enter. Advanced takes a beat to actually think — and on hard problems, that beat is the difference between a confident-sounding guess and a correct answer.
Understands
Reads the whole request — and the files and context around it — before committing to an approach.
Reasons
Works the problem through step by step, weighing options and edge cases instead of the first thing that fits.
Checks
Pressure-tests its own draft, catching mistakes a faster model would hand you unnoticed.
Answers
Delivers a clear, structured result — with the reasoning there when you want to see the working.
Use it when it truly counts.
Advanced uses 3× the tokens of Core, so it is built to be reached for deliberately — on the work where a better answer is genuinely worth it.
Reach for Advanced
- A gnarly bug or a large refactor across many files
- A contract, financial model or dense report to reason over
- Deep research that must hold together across many steps
- A high-stakes answer where being wrong is expensive
Core is plenty
- Everyday questions and quick drafts
- Summaries, rewrites and formatting
- Fast, grounded answers with web search
- Most day-to-day work — Core is fast and excellent
Switch tier per conversation — your history and files come with you. Reach for Advanced when a problem earns it, and drop back to Core the rest of the time.
Frontier-class intelligence.
On the hardest reasoning and coding work, Depaza Advanced performs in the same class as the leading US frontier models — landing between Opus 4.8 and Fable 5 in capability. That is the tier of intelligence you are reaching for, and it is why Advanced is metered at 3× tokens.
Same frontier tier — hosted entirely in the EU, and priced in tokens, not a per-seat premium.
A premium tier — on Pro+ and up.
Depaza Advanced unlocks on Pro+ (€119) and above. Its 3× token cost is why it lives on the power-user tiers — the plans with the weekly headroom to actually lean on it — not the entry plans.
Pro+
Where Advanced startsThe single-seat power-user tier — around 5× Pro's weekly headroom, built to actually live in Advanced.
Max
Our biggest weekly allowance and highest priority — for people who reach for the hardest tier every day.
Enterprise
Unlimited usage, SSO and a signed DPA — Advanced across your whole organisation.
Heavy stretch? A prepaid reserve pack (from €119) sits behind your plan and kicks in only after your weekly allowance runs out — no surprise bills, and unused credit carries over.
See the reserve pack →Why 3×?
Advanced is our most capable model and takes real work to run, so each token counts three times against your weekly budget. It is disclosed everywhere, and you stay in control — you choose when to use it, per conversation.
Questions, answered
Which model is Depaza Advanced?
Why does it cost 3× tokens?
Which plans include it?
Will it use up my weekly budget faster?
Can I switch back to Core mid-conversation?
Is my data used to train it?
Is it really EU-hosted?
Point your hardest problem at it.
Switch any conversation to Depaza Advanced and see what a model that thinks first can do. On Pro+ and above.