Changelog
Vynaris changelog
New features, models and improvements to Vynaris, newest first.
October 2026
- New
Pay with Monero
You can now pay for any top-up or plan in Monero (XMR) from the Billing page, with no card, name or billing address. Credit works the same as card-paid credit, including access to every model.
September 2026
- New
GLM-5.3 added to the hosted models
GLM-5.3-uncensored joins the privately hosted models you can select directly in a request, with 128K context and published per-token pricing.
- Improved
Simpler pricing and a smoother first payment
Prepaid top-ups now start at $20, and the pricing page and checkout were reworked so your first payment takes fewer steps.
August 2026
- Improved
Clear pricing and model pages for hosted models
Each hosted model now has its own page with its exact API ID, context size and per-token price. Model IDs were also simplified so they are easier to copy into your code.
- New
Privately hosted models
Vynaris now offers its own privately hosted Qwen and DeepSeek models with fewer refusals, for authorized security testing and model evaluation. Select one directly in your request; prompts and outputs are not retained.
July 2026
- Improved
A redesigned chat
Chat in your dashboard has a cleaner layout, and each reply shows which model answered and what it cost.
- New
Monthly plans
Choose a fixed monthly plan (S, L or XL) that adds API credit to your balance every month. Unused credit rolls over, and you can still top up at any time.
- New
Chat in your dashboard
Talk to Vynaris models directly from your dashboard without writing any code. Conversations are saved so you can pick up where you left off.
- Improved
Longer replies and steadier tool calls
The default reply length is now 8K tokens, so long answers are less likely to cut off. Tool calling from agents and coding tools now works more reliably.
- New
Guided setup and low-balance reminders
New accounts get a step-by-step welcome flow from funding your balance to sending your first request. The app now also reminds you when your credit is running low.
- Improved
Better automatic model choice
Setting the model to auto now puts answer quality first and works better across coding, analysis and image-based tasks. Requests also recover more gracefully when a model is temporarily unavailable.
- New
Model pricing table and cost calculator
Compare list prices for 30 major models on one page, then estimate your monthly bill with the calculator. An Enterprise option is also available for teams that want the gateway running on their own servers.
- New
Dashboard, API keys and setup guides
Sign in with email, GitHub or Google to get an API key, add credit and review your usage. New docs cover setup for Claude Code, Cursor, the OpenAI and Anthropic SDKs, Hermes, OpenClaw and Pi.
- New
Vynaris launches in early beta
Point any OpenAI-compatible client at one base URL and get a receipt with every response showing the model used, its cost and the direct price. Usage and ledger views let you check the math, and AI agents can sign up programmatically.