SpanveroHow it worksFind a modelCompare modelsPricing

What's new in Spanvero

New models every Monday — plus every feature we ship, logged honestly. Currently on v0.0.13.

v0.0.13 — Cost audits — now a three-tier ladder

July 9, 2026

The flat-fee cost audit grew into three published tiers: a $49 quick read (one focused question answered in writing, 2 business days), the unchanged $249 full written audit, and a $499 audit + prioritized implementation plan with one revision. Same honest flow — email intake, Stripe invoice only after you approve a written scope, and if we find no savings, the report says so.

v0.0.12 — State of Inference Costs — the monthly report

July 9, 2026

A new monthly report series at /reports, computed entirely from the append-only daily price log: the market floors at both ends of the window, every real price move with exact before→after numbers and dates, and a coverage-honest read of what actually changed. Free to cite with a link.

v0.0.11 — A clearer front door — and a way to reach us

July 9, 2026

The homepage now opens with a real model record — live, dated prices, three ways to run — and a one-click start. GPU pages link straight to the vendor that priced each rate, and the new /contact page puts a plain-text inbox one click from every footer.

v0.0.10 — Pick exactly how your rented GPU runs

June 18, 2026

The rent flow now offers three clear ways to run — Sleeps when idle (cheapest), Stays warm (instant), or a Dedicated Pod (full control) — with an honest side-by-side of speed and cost so anyone can choose.

v0.0.9 — Full-speed GPU sessions — rent a warm GPU

June 18, 2026

Rent a GPU on your own account and keep it warm for the session — every message is instant, full-speed, no cold-start wait. Or paste an endpoint you already have running. $0 markup, always your own account.

v0.0.8 — One-click GPU launch (Experimental)

June 18, 2026

Connect your own RunPod account and let Spanvero spin up a GPU for you. Bring-your-own-key stays the instant, reliable path — auto-launch is labeled Experimental while we make cold starts faster.

v0.0.7 — Chat history & search

June 18, 2026

Every conversation is saved, searchable, and reopens exactly where you left off — with the same model it was using.

v0.0.6 — Fusion — blend several models into one answer

June 18, 2026

Fusion (Mixture-of-Agents) asks up to three models the same question, then a synthesizer merges their drafts into a single answer. Runs entirely in your browser on your own keys, $0 markup.

Open the free advisor →

The weekly price index

A short email of real AI price moves, straight from the daily log — no hype. We're collecting the list now; the first issue goes out when it opens. Unsubscribe with one click.