
Berserk Models
In-house models across modalities.
Chat, text-to-image and text-to-video: the core model capabilities of Glacier AI.
Explore models ↗GLACIER AI · In-house models and multimodal tools
Built around our in-house Berserk models and paired with AI Tools,
covering text, images, speech and video, and turning model capability into real applications.

Chat, text-to-image and text-to-video: the core model capabilities of Glacier AI.
Explore models ↗
Six capability areas including video, text, images and speech. Find the right AI tool for each task.
Browse tools ↗
A single entry point to many models that simplifies protocol integration and project setup.
Learn about integration ↗Priced at 50% of the benchmark model's lowest market price of the day. Reference prices are updated or may change daily; the actual unit price is the day's quote.
Benchmarked against DeepSeek V4 flash
Benchmarked against GLM5.2
Benchmarked against MiniMax H3
Single-modal and cross-modal models
From chat to images and video, explore the in-house Berserk models.
Contact us →