End-to-end platform for developing your AI applications

No matter where you start, build and scale your AI with ByteCompute.

Book a Demo Contact Sales

Explore AI Models Directory

All categories and models you can try out and seamlessly integrate in your projects

openai/whisper-large-v3-turboStar Featured

Whisper Large-Turbo

automatic-speech-recognition

A weakly supervised pre-trained version of the Whisper model, optimized for high-speed Automatic Speech Recognition (ASR) and speech translation. By significantly reducing the number of decoder layers to 4 while maintaining the robust large-v3 encoder architecture, this 'Turbo' variant offers an 8.8x speedup compared to large-v3 with minimal degradation in Word Error Rate (WER). It is specifically designed as a high-efficiency alternative for low-latency production environments.

Memory 2 GBSetting 809 M$0.00066/MINUTE
Qwen3.5-122B-A10B-FP8

Qwen3.5 122B

TEXT

The flagship Mixture-of-Experts (MoE) model from the Qwen3.5 series, featuring 122B total and 10B active parameters. This unified vision-language foundation excels in multimodal reasoning, complex coding, and native 'thinking mode' tasks. Utilizing fine-grained FP8 quantization, it offers exceptional throughput and reduced VRAM footprint on H100/L40S GPUs, while supporting a massive 262K context window for long-horizon agentic applications.

Memory 122GBSetting 122BINPUT $0.400/1M TOKENS ;OUTPUT $3.200/1M TOKENS
Lightricks/LTX-2.3Star Featured

LTX-2.3

VIDEO

A state-of-the-art Diffusion Transformer (DiT) foundation model with 22 billion parameters. Unlike traditional video models, LTX-2 is natively designed for synchronized audio-video generation within a single unified latent space. It excels at maintaining temporal consistency and high-fidelity motion, making it a powerful backend for creative AI pipelines that require seamless audiovisual coherence.

Memory 22GBSetting 22B$0.110/SECOND
openai/whisper-large-v3

Whisper Large V3

automatic-speech-recognition

The Whisper large-v3 is a pre-trained model for Automatic Speech Recognition (ASR) and speech translation. It features a robust Transformer encoder-decoder architecture designed for state-of-the-art accuracy across a wide range of languages and audio conditions.

Memory 2GBSetting 809M$0.0015/MINUTE
flux2-klein-4bStar Featured

FLUX Klein

IMAGE

The FLUX.2 [klein] model family are our fastest image models to date. FLUX.2 [klein] unifies generation and editing in a single compact architecture, delivering state-of-the-art quality with end-to-end inference in as low as under a second.

Memory 13GBSetting 4B$0.003/IMAGE
gemma-4-31B-itStar Featured

Gemma 4 31B

TEXT

gemma-4-31B-it is a 30.7-billion parameter dense multimodal model. It is built using the same research infrastructure as Gemini 3, offering frontier-level performance for reasoning, agentic workflows, and high-accuracy coding. It natively supports text and image inputs with a massive 256K context window.

Memory /Setting 31bINPUT $0.140/1M TOKENS; OUTPUT $0.40/1M TOKENS
Qwen3.6-35B-A3B

Qwen3.6-35B-A3B

TEXT

Qwen3.6-35B-A3B is an efficient mixture-of-experts (MoE) model with 35B total parameters and only 3B active. It delivers strong agentic coding performance, significantly outperforming Qwen3.5-35B-A3B and competing with larger dense models like Qwen3.5-27B and Gemma4-31B. Supporting both multimodal thinking and non-thinking modes, it is a highly versatile open-source model, now available on Qwen Studio, via API, and as open weights.

Memory ~20GB (4-bit) / ~40GB (INT8) / ~80GB (FP16)Setting 35B (MoE, 3B active)INPUT $0.161/1M TOKENS; OUTPUT$0.965/1M TOKENS
paddleocr

PaddleOCR

OCR

PaddleOCR Service is a synchronous HTTP API for extracting text and document structure from images. Each request uploads a single image, the server runs inference, and the result is returned in the same response — no polling required.

Memory -Setting --
Qwen3.8-27B-FP8Star Featured

Qwen3.8-27B-FP8

TEXT

Qwen3.8-27B-FP8 is a 27B dense vision-language model deployed with FP8 quantization. This endpoint supports text, image and video inputs, text output, and reasoning control. The Bytecompute deployment supports up to 262,144 context tokens.

Memory ~31GB weights; 40GB+ VRAM recommendedSetting 27BINPUT $0.15/1M TOKENS ;OUTPUT $2.0/1M TOKENS

Ready to Accelerate AI in Your Organization?

Contact our sales team to discuss your enterprise needs and deployment options.

Get Started