We’ve got some exciting new models for you! We worked hard to bring something for all paid tiers—whether you love deep storytelling, anime-style interactions, or long, immersive chats. Here’s what’s new!

💎 I’m All In Models – For the Ultimate Experience

Deepseek V3 671B

 We love to push limits and see what’s possible. This time, we're taking a huge leap—introducing a model that's 5x larger than our biggest model, PixelXL, making it the most advanced AI we've ever released. Expect deeper conversations, smarter responses, and more immersive roleplays.

⚠ Due to its size, it may respond slower—but the experience is worth it! Available until at least April 15th.

Euryale 70B

 If you loved Stheno, you’ll love Euryale even more! This model is better at remembering details, sticking to prompts (better continuity), and keeping your roleplay engaging. Great for those who want more control over their character interactions.

PixelXL 132B – Now with 16K Context Memory!

 PixelXL has already become a fan favorite, and now it’s even better!

• We’ve doubled its memory—meaning it can now remember more details from past messages.
• It’s great for long conversations and can handle multiple characters in one scene with ease.

🎀 True Supporters

Shimizu 24B

 Perfect for deep, meaningful roleplays that is multilingual, supporting French, German, Spanish, Russian, Chinese, Korean, and Japanese. If you love well-developed stories, this is a great choice.

🌸 Get A Taste

Stheno 8B

 A great mid-sized model that’s fun, engaging, and keeps roleplays dynamic. If you want a mix of creativity and speed, this is an excellent choice.

💭 What’s Up with All These Numbers?

The number next to the model (e.g., 671B, 132B) is the number of AI parameters in billions.

• Bigger models = Smarter AI, more knowledge, and better responses.
• Larger models can also speak more languages and handle more complex interactions.
• Our biggest model is 80x larger than our default AI (and 80x more expensive to run 😅).

🧠 What is 16K context memory?

This tells you how many words (tokens) the model can remember.

A 16K context can store about 85 messages from your chat history for each generation. This means your characters will have better memory and continuity!

We’re so excited to hear what you think—which model are you trying first?

To change Model access, go to "Generation Settings" from the conversation menu on the chat page: