Referral credits, price drops, and more!

Published 2025-03-14

Thank you to everyone for supporting us so early! We're shipping referrals + referral credits: if you refer a friend, each of you gets $5 in credits as soon as they spend $10. We also dropped the price of input tokens for DeepSeek R1 to match our industry providers, added an improved API endpoint, and shipped a bundle of other features:

• Referral credits
• Price drops for DeepSeek R1
• New and improved API endpoint
• New payment provider support
• Bugfixes, performance improvements, and UI tweaks
• Future plans

We've added referral links and referral credits to Synthetic! If you go to synthetic.new/referrals, you can see your unique referral link: give that to your friends, followers, lovers, and enemies, and when each of them signs up and spends their first $10, you'll each get an extra $5 in credits to use on Synthetic. It's not exactly a free tier, but it's close!

For always-on models, we aim to price match our industry providers. For DeepSeek models, that's Together.ai. They dropped the cost of R1 input tokens to $3 per million tokens (output tokens remain $7/million), and we quietly dropped our prices last month as well to match. Now we're telling you about it!

Our old API endpoint was https://synthetic.new/api/openai/v1. That endpoint will continue to work exactly how it always has... But, there was a problem with it: because we used Cloudflare proxy for our DNS hosting, they enforced a 100-second timeout for responses. For models that take a long time to boot, you could see DNS-related timeouts when using the API.

We've added a new, unproxied DNS entry, and a new API endpoint that won't have that 100s timeout: https://api.synthetic.new/v1. You'll be able to see it in your API settings page, so go ahead and update any API clients to the new and improved API base URL!

We use Stripe to process our payments, meaning we never see your credit card directly. However, we heard from you that some people wanted more anonymity, so we're happy to announce we now accept crypto payments: we accept USDC on Ethereum, Solana, and Polygon, and USDP on Ethererum and Solana. One caveat, though, is that currently we only support crypto payments as manual payments (i.e. auto-refill still requires a credit card).

We've also added support for Amazon Pay and Apple Pay!

Usually this section is a little boring, but we do have some more-notable changes this time around: we render DeepSeek-style <think> sections in a nice little UI bubble (you'll always be able to see all of the thinking tokens — we don't hide them!), and we added support for rendering tables and certain math symbols in the UI.

If you made it here, thank you so much for your support and being a part of our journey!

We're hard at work on more improvements, including:

• Multi-modal image input, so you can chat with LLMs about pictures. We're hard at work on this: expect an update in the next newsletter!
• Better ways to share links to models you like, and optionally share chat outputs that are fun or interesting.
• Infrastructure improvements for faster, cheaper on-demand models, in addition to the always-on LoRAs.
• Dark mode! We've had a lot of requests for this one and it's on our radar. We're working with a designer to clean up the UI and figure out a good colorscheme.
• Community Discord! This has been a to-do for a bit for us, but we promise it's actually coming.
• More backend and UI improvements. Maybe not the most exciting line item, but we're always bugfixing.

If you have any thoughts or feedback, please continue to reach out at [email protected]. We appreciate all the emails we've gotten so far!

— Matt & Billy