Model Library
Browse and deploy state-of-the-art AI models through the DEVUP Gateway.
Model Library
Browse and deploy state-of-the-art AI models through the DEVUP Gateway.
Browse and deploy state-of-the-art AI models through the DEVUP Gateway.


Claude Fable 5 was Anthropic's most capable widely released model, built for demanding reasoning and long-horizon agentic work. Two things set it apart from everything else in this catalogue. Thinking is permanently on — there is no parameter that disables it, only an effort dial that sets how deep it goes. And it ships with safety classifiers that can decline a request, returning a successful response carrying a refusal rather than an error, which means any integration must inspect why a response ended rather than assuming a 200 means an answer. It holds a million-token context with a 128,000-token output ceiling, and reads text and images.


Claude Sonnet 5 pairs a million-token context window with the fastest comparative latency in its generation, which is what makes it the working default rather than an economy option. It matches the Opus tier on window size and output ceiling while answering noticeably quicker, and adaptive thinking is on by default with five effort levels controlling how deep it goes. One integration detail separates it from most models in this catalogue: sampling parameters are rejected outright. Send a temperature, a top-p, or a top-k at anything other than its default and the request returns an error rather than quietly ignoring it — so output variation is shaped through effort and prompting instead.


MiniMax-M3 is built around one component: a sparse attention operator that MiniMax published as a standalone project with its own technical report. The numbers it produces are measured against the previous generation at a million tokens — nine times faster prefill, fifteen times faster decode, and per-token compute cut to one twentieth. Multimodality was not added afterwards; mixed-modality training ran from the first step across text, image, and video. Four hundred and twenty-eight billion parameters with twenty-three active, sixty layers, a 524,288-token window, and a thinking parameter that can let the model decide for itself when deliberation is worth it.

Turn one portrait photo into a talking video. P-Video-Avatar makes the person in your image speak your script with natural lip sync, in one of 30 voices and 10 languages, in 720p or 1080p. Made for spokesperson clips, product explainers, customer messages and multilingual content.


The 1080p high-fidelity mode in PixVerse renders videos with significantly enhanced sharpness and visual clarity, capturing intricate details and providing a crisp, professional-grade quality suitable for more polished projects.


PixVerse's 720p resolution offers a fast and reliable option for generating standard HD videos, ideal for quick previews and social media content where generation speed is prioritized over maximum detail.


PixVerse V6 turns a text prompt into a video of up to 15 seconds with native audio: dialogue, sound effects and ambient sound generated with the picture. Four resolutions from 360p to 1080p, eight aspect ratios from vertical 9:16 to cinematic 21:9, and an audio switch that lowers the price when you bring your own soundtrack.


PixVerse V6 brings a still image to life: your picture becomes the first frame of a clip of up to 15 seconds, from 360p to 1080p, with native audio (dialogue, sound effects and ambient sound generated with the picture). The video follows the shape of your image, and an audio switch lowers the price when you bring your own soundtrack.

A new-generation professional-grade multimodal video creation model developed, supports video generation with multimodal reference inputs including images, videos and audio.

ByteDance's Seedance 1.5 Pro is a professional video model using V2A native generation for integrated, synced audio-visual output, enhancing efficiency of professional video creation.


SDXL Turbo is Stability AI's real-time text-to-image model: SDXL 1.0 distilled with Adversarial Diffusion Distillation, making high-quality images in 1 to 4 steps. Priced by size and steps, up to 4 images per request.


P-Image-Edit is Pruna AI's real-time image editing model: insertions, removals, transformations and style changes in under a second according to Pruna, from up to 5 input images, at seven preset formats or any size up to 1440 pixels per side.


P-Image is Pruna AI's real-time text-to-image model: a state-of-the-art image in under one second according to Pruna, with exact prompt adherence and controlled text, seven preset formats or any size up to 1440 pixels per side.


ClarityAI/creative is an AI-powered image upscaler that enhances details, adds realism, and creatively modifies images to improve their quality and visual appeal.

Seedream 4.5 is ByteDance Seed's scaled-up successor to Seedream 4.0: better on every text-to-image and editing axis ByteDance measures, with stronger editing consistency and dense-text typography, at 2K to 4K and one price per image.

Seedream 4.0 is ByteDance Seed's image model that unifies generation and editing in one diffusion transformer: images up to 4096×4096 at any format, text rendering, and edits from one sentence, at one price per image whatever the size.


Realtime TTS 2.0 is a low-latency text-to-speech model with natural language steering, allowing you to control tone and emotion directly in the prompt (e.g., “[be happy and upbeat] Hello!”). It supports cross-lingual voices and multiple languages, enabling the same voice to speak consistently across different languages. This is an early access preview ahead of full launch, with ongoing improvements to voice quality and steering.


Fast multilingual text-to-speech model by Inworld AI with 130+ preset voices across 15 languages. Supports voice cloning, word-level timestamps, and streaming. Optimized for low-latency applications with <130ms time-to-first-audio.


High-quality multilingual text-to-speech model by Inworld AI with 130+ preset voices across 15 languages. Supports voice cloning, word-level timestamps, and streaming. Optimized for natural, expressive speech with <250ms time-to-first-audio.

Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers comparable quality to larger models while being significantly faster and more cost-efficient. With Apache-licensed weights, Kokoro can be deployed anywhere from production environments to personal projects.


Orpheus TTS is a state-of-the-art, Llama-based Speech-LLM designed for high-quality, empathetic text-to-speech generation. This model has been finetuned to deliver human-level speech synthesis, achieving exceptional clarity, expressiveness, and real-time streaming performances.


HiggsAudioV2.5 is a high-quality neural text-to-speech (TTS) model designed for natural-sounding voice generation across a wide range of use cases. It focuses on clarity, stable prosody, and consistent pacing, making it suitable for both short prompts and longer narration.


09/04 🔥 Introducing Chatterbox Multilingual in 23 Languages! We're excited to introduce Chatterbox and Chatterbox Multilingual, Resemble AI's production-grade open source TTS models. Chatterbox Multilingual supports Arabic, Danish, German, Greek, English, Spanish, Finnish, French, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Dutch, Norwegian, Polish, Portuguese, Russian, Swedish, Swahili, Turkish, Chinese out of the box. Licensed under MIT, Chatterbox has been benchmarked against leading closed-source systems like ElevenLabs, and is consistently preferred in side-by-side evaluations.


Nemotron 3.5 Content Safety is a four-billion-parameter moderator that can explain itself. It classifies a prompt, an optional image, and an optional response across a twenty-three-category taxonomy — and in think mode it emits an auditable reasoning trace before the verdict, which is what makes a blocked decision defensible rather than merely enforced. It also accepts your own safety policy in place of the default taxonomy. Its twelve languages were not translated but culturally adapted, addressing a measured problem: models produce more unsafe content in non-English languages because safety training data lacks cultural nuance. It runs on a single L4.