AI models

Our small fast model designed for creativity, roleplay and general tasks

Our latest and highest-quality model designed for creativity, roleplay and general tasks with deep thinking ability

Our latest and highest-quality model designed for creativity, roleplay and general tasks

Best quality, optimized for roleplay

DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on output from a 552B-parameter backbone, an asymmetric split that keeps per-token compute low relative to the model's total size

Excellent, optimized for roleplay

DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.

Our latest TTS model for generating high-quality Kurdish speech.

Our latest ASR model for accurate Kurdish speech recognition.

Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation

Chatterbox is a family of three state-of-the-art, open-source text-to-speech models by Resemble AI.

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters.

DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.

DeepSeek V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes.

MiniMax-M2.5 is a SOTA large language model designed for real-world productivity.

Engineered to deliver frontier reasoning and agentic capabilities with exceptional efficiency.

Designed for powerful reasoning, agentic tasks, and versatile developer use cases.

Excellent model, can be used for general tasks.