NovitaAI Provider

Ming Image 0.1 Design is a text-to-image model from inclusionAI aimed at graphic-design output, with an emphasis on legible text rendering inside the generated image. It generates from a prompt only and does not accept reference images. Output format can be requested as PNG, JPEG, or WebP. Image dimensions are chosen by the model rather than by the request, so explicit sizes and aspect ratios are rejected instead of silently reshaped.

Ming Image 0.1 Design Layer is an image-to-image model from inclusionAI that decomposes a flattened design image into separate RGBA layers, such as a background layer and foreground elements, and returns one image per layer. It requires exactly one reference image and a prompt describing the layer plan. Output format can be requested as PNG or WebP. Output dimensions follow the input image, so explicit sizes and aspect ratios are rejected instead of silently reshaped.

Ling-3.0-flash is the latest-generation cost-effective model in the Ling series, with 124B total parameters, 5.1B activated parameters, a native 256K context window extendable up to 1M. Compared to the previous flash model, Ling-3.0-flash delivers significantly enhanced stability in long-horizon tasks, markedly higher tool calling accuracy, and substantially improved compatibility with common Harness environments. In coding tasks, Ling-3.0-flash features greatly optimized spatial understanding, capable of constructing physical scene grids and relative positions. Additionally, thanks to its hybrid Reasoning mode, task success rates across varying difficulty levels have all improved.

Tencent: Hy4 preview is a mixture-of-experts model from Tencent, with 49B active parameters out of 770B total. It is designed for coding agents, complex tool-use workflows, and productivity tasks that require planning, context continuity, and sustained multi-step execution.

Kimi K3 is Kimi’s most capable model to date, with 2.8 trillion parameters. Built on Kimi Delta Attention, a hybrid linear attention mechanism, and Attention Residuals, it offers native visual understanding and a 1M-token context window for frontier intelligence scenarios such as software engineering, knowledge work, and deep reasoning.

Gemma 4 31B IT is an instruction-tuned multimodal open model from Google DeepMind with a 256K context window, designed for reasoning, coding, agentic workflows, and image understanding. It accepts text and image inputs and generates text output, with native tool calling and configurable thinking modes for long-context assistant experiences.

Macaron-V1-Venti is a 748B-parameter flagship model in the Macaron-V1 family, built for personal intelligence, tool use, coding workflows, and code-native Generative UI. The model uses a Mixture of LoRA (MoL) architecture on top of GLM-5.2, consisting of a 744B-parameter base model and four 1B-parameter LoRA specialists. The specialists cover chat, personal-agent tasks, coding, and GenUI, with an L0 router selecting the most suitable specialist for each new user request.

Macaron-V1-Tall is the efficient model of the Macaron-V1 family, built on Qwen3.6-35B-A3B with four specialist LoRA adapters for Chat, Agent, Coding, and GenUI. It follows the same routed Mixture of LoRA (MoL) design as Macaron-V1-Venti, bringing Macaron’s personal-intelligence, coding, tool-use, and code-native Generative UI capabilities to a smaller and more accessible deployment profile.

Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimodal input (text and image) and multilingual output (text and code) across 12 supported languages. Designed for assistant-style interaction and visual reasoning, Scout uses 16 experts per forward pass and features a context length of 10 million tokens, with a training corpus of ~40 trillion tokens. Built for high efficiency and local or commercial deployment, Llama 4 Scout incorporates early fusion for seamless modality integration. It is instruction-tuned for use in multilingual chat, captioning, and image understanding tasks. Released under the Llama 4 Community License, it was last trained on data up to August 2024 and launched publicly on April 5, 2025.

Ling-3.0-flash-Fin is a finance-enhanced MoE model built on Ling-3.0-flash, with 124 billion total parameters and approximately 5.1 billion activated parameters. Designed for real-world investment workflows, it is optimized for complex multi-step tasks and long-horizon planning and execution. With a relatively small active parameter footprint, it delivers competitive financial performance while maintaining strong general capabilities in reasoning, coding, and mathematics.

Ling-3.0-Flash-Sante is a Mixture-of-Experts (MoE) model with enhanced capabilities across health and medicine. Built on Ling-3.0-Flash, it has 124 billion total parameters and activates approximately 5.1 billion parameters per token. The model excels in medical knowledge reasoning, clinical safety, evidence-based retrieval, and long-horizon medical tasks, while retaining strong general capabilities in reasoning, coding, and agentic tasks.

Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at a fraction of the compute cost. Supports multimodal input including text, images, and video (up to 60s at 1fps). Features a 256K token context window, native function calling, configurable thinking/reasoning mode, and structured output support. Released under Apache 2.0.