This website uses cookies to anonymously analyze website traffic using Google Analytics.
AcceptDecline

Together GPU Clusters featuring NVIDIA GB200, H200, and H100. Request your cluster today

The AI Acceleration AccelerationCloud

Train, fine-tune-and run inference on AI models blazing fast, at low cost, and at production scale.

Start building nowContact sales

Trusted by

200+ generative AI models

Build with open-source and specialized multimodal models for chat, images, code, and more. Migrate from closed models with OpenAI-compatible APIs.

All
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Try now
together.ai

Image

FLUX.1 [dev]

12 billion parameter rectified flow transformer capable of generating images from text descriptions.

TRY
THIS MODEL

Image

FLUX.1 Canny [dev]

12 billion parameter rectified flow transformer capable of generating an image based on a text description while following the structure of a given input image.

TRY
THIS MODEL

Image

FLUX.1 Depth [dev]

12 billion parameter rectified flow transformer capable of generating an image based on a text description while following the structure of a given input image.

TRY
THIS MODEL

Image

FLUX.1 Redux [dev]

Adapter for FLUX.1 models enabling image variation, refining input images, and integrating into advanced restyling workflows.

TRY
THIS MODEL

Chat

Typhoon 1.5 8B Instruct

Instruct Thai large language model with 8 billion parameters based on Llama3-8B.

TRY
THIS MODEL

Chat

Typhoon 1.5X 70B-awq

Thai language 70B instruct model rivaling GPT-4-0612; optimized for RAG, constrained generation, and reasoning tasks.

TRY
THIS MODEL

Chat

Llama 3.3 70B

The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model is optimized for multilingual dialogue use cases and outperform many of the available open source and closed chat models on common industry benchmarks.

TRY
THIS MODEL

Image

FLUX1.1 [pro]

Premium image generation model by Black Forest Labs.

TRY
THIS MODEL

Chat

Gemma-2 Instruct (27B)

Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models.

TRY
THIS MODEL

Chat

Qwen 2.5 72B

Powerful decoder-only models available in 7B and 72B variants, developed by Alibaba Cloud's Qwen team for advanced language processing.

TRY
THIS MODEL

Chat

Llama 3.1 Nemotron 70B Instruct

This LLM is customized by NVIDIA to improve the helpfulness of LLM generated responses to user queries.

TRY
THIS MODEL

Language

Mixtral-8x22B

The Mixtral-8x22B Large Language Model (LLM) is a pretrained generative Sparse Mixture of Experts.

TRY
THIS MODEL

Chat

DBRX-Instruct

DBRX Instruct is a mixture-of-experts (MoE) large language model trained from scratch by Databricks. DBRX Instruct specializes in few-turn.

TRY
THIS MODEL

Chat

Deepseek-67B

Trained from scratch on a vast dataset of 2 trillion tokens in both English and Chinese.

TRY
THIS MODEL

Chat

Arctic-Instruct

Arctic is a dense-MoE Hybrid transformer architecture pre-trained from scratch by the Snowflake AI Research Team.

TRY
THIS MODEL

Chat

Striped Hyena Nous

A hybrid architecture composed of multi-head, grouped-query attention and gated convolutions arranged in Hyena blocks, different from traditional decoder-only Transformers

TRY
THIS MODEL

Chat

Llama 3.2 11B Free

Free endpoint to try Llama 3.2 11B.

TRY
THIS MODEL

Free

Chat

Llama 3.2 90B

The Llama 3.2-Vision collection features multimodal LLMs (11B and 90B) optimized for visual recognition, image reasoning, captioning, and answering image-related questions.

TRY
THIS MODEL

Chat

Llama 3.1 8B

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pre-trained and instruction tuned generative models in 8B, 70B and 405B sizes, that outperform many available open source and closed chat models on common industry benchmarks.

TRY
THIS MODEL

Chat

Llama 3.2 11B

The Llama 3.2-Vision collection features multimodal LLMs (11B and 90B) optimized for visual recognition, image reasoning, captioning, and answering image-related questions.

TRY
THIS MODEL

Image

FLUX.1 [schnell] Free

Free endpoint for the SOTA open-source image generation model by Black Forest Labs.

TRY
THIS MODEL

Free

Image

FLUX.1 [schnell]

Fastest available endpoint for the SOTA open-source image generation model by Black Forest Labs.

TRY
THIS MODEL

Chat

Llama 3.1 405B

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pre-trained and instruction tuned generative models in 8B, 70B and 405B sizes, that outperform many available open source and closed chat models on common industry benchmarks.

TRY
THIS MODEL

Chat

Qwen QwQ 32B Preview

Experimental research model by Alibaba's Qwen team focused on enhancing AI reasoning capabilities.

TRY
THIS MODEL

Embeddings

UAE-Large v1

An universal English sentence embedding model by WhereIsAI. Its embedding dimension is 1024, it takes up to 512 context length.

TRY
THIS MODEL

Chat

Qwen 2

A transformer-based decoder-only language model pre-trained on a large amount of data. In comparison with the previously released Qwen

TRY
THIS MODEL

Language

01-AI Yi

The Yi series models are large language models trained from scratch by developers at 01.AI

TRY
THIS MODEL

Code

Qwen 2.5 Coder 32B Instruct

SOTA code LLM with advanced code generation, reasoning, fixing, and support for up to 128K tokens.

TRY
THIS MODEL

Chat

Mixtral 8x7B

The Mixtral-8x7B Large Language Model (LLM) is a pretrained generative Sparse Mixture of Experts.

TRY
THIS MODEL

Embeddings

M2-BERT 80M 32K Retrieval

An 80M checkpoint of M2-BERT, pretrained with sequence length 32768, and it has been fine-tuned for long-context retrieval.

TRY
THIS MODEL

Image

Stable Diffusion XL 1.0

A text-to-image generative AI model that excels at creating 1024x1024 images.

TRY
THIS MODEL

Chat

Mistral Instruct

instruct fine-tuned version of Mistral-7B-v0.1

TRY
THIS MODEL

Language

Nexus Raven

NexusRaven is an open-source and commercially viable function calling LLM that surpasses the state-of-the-art in function calling capabilities.

TRY
THIS MODEL

Chat

Nous Capybara

First Nous collection of dataset and models made by fine-tuning mostly on data created by Nous in-house

TRY
THIS MODEL

Chat

Code Llama Instruct

Code Llama is a family of large language models for code based on Llama 2 providing infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks.

TRY
THIS MODEL

Language

RedPajama-INCITE Instruct

Designed for few-shot prompts, fine-tuned over the RedPajama-INCITE-Base-7B-v1 base model.

TRY
THIS MODEL

Chat

vicuna-v1-5-16k

Vicuna is a chat assistant trained by fine-tuning Llama 2 on user-shared conversations collected from ShareGPT.

TRY
THIS MODEL

Chat

Nous Hermes Llama-2

Nous-Hermes-Llama2-13b is a state-of-the-art language model fine-tuned on over 300,000 instructions.

TRY
THIS MODEL

Chat

Wizard LM

This model achieves a substantial and comprehensive improvement on coding, mathematical reasoning and open-domain conversation capacities

TRY
THIS MODEL

Embeddings

BGE-Large-EN v1.5

BAAI general embedding - large, english - model v1.5. FlagEmbedding can map any text to a low-dimensional dense vector which can be used for tasks like retrieval, classification, clustering, or semantic search. And it also can be used in vector databases for LLMs.

TRY
THIS MODEL

Embeddings

M2-BERT 80M 2K Retrieval

An 80M checkpoint of M2-BERT, pretrained with sequence length 2048, and it has been fine-tuned for long-context retrieval.

TRY
THIS MODEL

Embeddings

M2-BERT 80M 8K Retrieval

An 80M checkpoint of M2-BERT, pretrained with sequence length 8192, and it has been fine-tuned for long-context retrieval.

TRY
THIS MODEL

Code

WizardCoder Python v1.0

This model empowers Code LLMs with complex instruction fine-tuning, by adapting the Evol-Instruct method to the domain of code.

TRY
THIS MODEL

Embeddings

BGE-Base-EN v1.5

BAAI general embedding - base, english - model v1.5. FlagEmbedding can map any text to a low-dimensional dense vector which can be used for tasks like retrieval, classification, clustering, or semantic search. And it also can be used in vector databases for LLMs.

TRY
THIS MODEL

Chat

LLaMA-2-7B-32K-Instruct

Extending LLaMA-2 to 32K context, built with Meta's Position Interpolation and Together AI's data recipe and system optimizations, instruction tuned by Together

TRY
THIS MODEL

Language

Mistral

7.3B parameter model that outperforms Llama 2 13B on all benchmarks, approaches CodeLlama 7B performance on code, Uses Grouped-query attention (GQA) for faster inference and Sliding Window Attention (SWA) to handle longer sequences at smaller cost

TRY
THIS MODEL

Code

Code Llama Python

Code Llama is a family of large language models for code based on Llama 2 providing infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks.

TRY
THIS MODEL

Chat

Vicuna v1.5

Vicuna is a chat assistant trained by fine-tuning Llama 2 on user-shared conversations collected from ShareGPT.

TRY
THIS MODEL

Language

Phind Code LLaMA v2

Phind-CodeLlama-34B-v1 trained on additional 1.5B tokens high-quality programming-related data proficient in Python, C/C++, TypeScript, Java, and more.

TRY
THIS MODEL

Language

LLaMA-2

Language model trained on 2 trillion tokens with double the context length of Llama 1. Available in three sizes: 7B, 13B and 70B parameters

TRY
THIS MODEL

Code

Code Llama

Code Llama is a family of large language models for code based on Llama 2 providing infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks.

TRY
THIS MODEL

Language

LLaMA-2-32K

Extending LLaMA-2 to 32K context, built with Meta's Position Interpolation and Together AI's data recipe and system optimizations.

TRY
THIS MODEL

Chat

Chronos Hermes

This model is a 75/25 merge of Chronos (13B) and Nous Hermes (13B) models resulting in having a great ability to produce evocative storywriting and follow a narrative.

TRY
THIS MODEL

Rerank

Salesforce LlamaRank

Salesforce Research's proprietary fine-tuned rerank model with 8K context, outperforming Cohere Rerank for superior document retrieval.

TRY
THIS MODEL

Chat

Platypus2 Instruct

An instruction fine-tuned LLaMA-2 (70B) model by merging Platypus2 (70B) by garage-bAInd and LLaMA-2 Instruct v2 (70B) by upstage.

TRY
THIS MODEL

Language

WizardLM v1.0 (70B)

This model achieves a substantial and comprehensive improvement on coding, mathematical reasoning and open-domain conversation capacities.

TRY
THIS MODEL

Chat

MythoMax-L2

MythoLogic-L2 and Huginn merge using a highly experimental tensor type merge technique. The main difference with MythoMix is that I allowed more of Huginn to intermingle with the single tensors located at the front and end of a model

TRY
THIS MODEL

Chat

Qwen-Chat

7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B-Chat is a large-model-based AI assistant, which is trained with alignment techniques.

TRY
THIS MODEL

Language

Qwen

7B-parameter version of the large language model series, Qwen (abbr. Tongyi Qianwen), proposed by Aibaba Cloud. Qwen-7B is a Transformer-based large language model, which is pretrained on a large volume of data, including web texts, books, codes, etc.

TRY
THIS MODEL

Image

FLUX.1 [pro]

First generation premium image generation model by Black Forest Labs.

TRY
THIS MODEL

Chat

RedPajama-INCITE Chat

Chat model fine-tuned using data from Dolly 2.0 and Open Assistant over the RedPajama-INCITE-Base-7B-v1 base model.

TRY
THIS MODEL

Embeddings

BERT

Pretrained model on English language using a masked language modeling (MLM) objective. The embedding dimension is 768, and the number of model parameters is 110M. This model is uncased: it does not make a difference between english and English.

TRY
THIS MODEL

Language

RedPajama-INCITE

Base model that aims to replicate the LLaMA recipe as closely as possible (blog post).

TRY
THIS MODEL

Embeddings

Sentence-BERT

This is a sentence-transformers model: It maps sentences & paragraphs to a 768 dimensional dense vector space and was designed for semantic search. It has been trained on 500K (query, answer) pairs from the MS MARCO dataset. Its embedding dimension is 768 with 512 max context length.

TRY
THIS MODEL

Language

GPT-JT-Moderation

This model can be used to moderate other chatbot models. Built using GPT-JT model fine-tuned on Ontocord.ai's OIG-moderation dataset v0.1.

TRY
THIS MODEL

Language

GPT-JT

Fork of GPT-J instruction tuned to excel at few-shot prompts (blog post).

TRY
THIS MODEL

Chat

GPT-NeoXT-Chat-Base

Chat model fine-tuned from EleutherAI’s GPT-NeoX with over 40 million instructions on carbon reduced compute.

TRY
THIS MODEL

Language

LLaMA

An auto-regressive language model, based on the transformer architecture. The model comes in different sizes: 7B, 13B, 33B and 65B parameters.

TRY
THIS MODEL

Chat

Falcon Instruct

Falcon-40B-Instruct is a causal decoder-only model built by TII based on Falcon-40B and finetuned on a mixture of Baize.

TRY
THIS MODEL

Language

Falcon

Falcon-40B is a causal decoder-only model built by TII and trained on 1,000B tokens of RefinedWeb enhanced with curated corpora.

TRY
THIS MODEL

Llama 3.1 70B

The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pre-trained and instruction tuned generative models in 8B, 70B and 405B sizes, that outperform many available open source and closed chat models on common industry benchmarks.

TRY
THIS MODEL

Chat

Pythia-Chat-Base

Chat model based on EleutherAI’s Pythia-7B model, and is fine-tuned with data focusing on dialog-style interactions.

TRY
THIS MODEL

End-to-end platform for the full generative AI lifecycle

Leverage pre-trained models, fine-tune them for your needs, or build custom models from scratch. Whatever your generative AI needs, Together AI offers a seamless continuum of AI compute solutions to support your entire journey.

  • Inference

    The fastest way to launch AI models:

    • ✔ Serverless or dedicated endpoints

    • ✔ Deploy in enterprise VPC

    • ✔ SOC 2 and HIPAA compliant

  • Fine-Tuning

    Tailored customization for your tasks

    • ✔ Complete model ownership

    • ✔ Fully tune or adapt models

    • ✔ Easy-to-use APIs

    • Full Fine-Tuning
    • LoRA Fine-Tuning

  • GPU Clusters

    Full control for massive AI workloads

    • ✔ Accelerate large model training

    • ✔ GB200, H200, and H100 GPUs

    • ✔ Pricing from $1.75 / hour

Run
models

Train

Models

Speed, cost, and accuracy. Pick all three.

SPEED RELATIVE TO VLLM

4x FASTER

LLAMA-3 8B AT FULL PRECISION

400 TOKENS/SEC

COST RELATIVE TO GPT-4o

11x lower cost

Why Together Inference

Powered by the Together Inference Engine, combining research-driven innovation with deployment flexibility.

  • accelerated by cutting edge research

    Transformer-optimized kernels: our researchers' custom FP8 inference kernels, 75%+ faster than base PyTorch

    ‍

    Quality-preserving quantization: accelerating inference while maintaining accuracy with advances such as QTIP

    ‍

    Speculative decoding: faster throughput, powered by novel algorithms and draft models trained on RedPajama dataset

  • Flexibility to choose a model that fits your needs

    Turbo: ​​Best performance without losing accuracy

    ‍

    Reference: Full precision, available for 100% accuracy

    ‍

    Lite: Optimized for fast performance at the lowest cost

  • Available via Dedicated instances and serverless API

    Dedicated instances: fast, consistent performance, without rate limits, on your own single-tenant NVIDIA GPUs

    ‍

    Serverless API: quickly switch from closed LLMs to models like Llama, using our OpenAI compatible APIs

Control your IP.
‍Own your AI.

Fine-tune open-source models like Llama on your data and run them on Together Cloud or in a hyperscaler VPC. With no vendor lock-in, your AI remains fully under your control.

together files upload acme_corp_customer_support.jsonl
  
{
  "filename" : "acme_corp_customer_support.json",
  "id": "file-aab9997e-bca8-4b7e-a720-e820e682a10a",
  "object": "file"
}
  
  
together finetune create --training-file file-aab9997-bca8-4b7e-a720-e820e682a10a
--model together compute/RedPajama-INCITE-7B-Chat


together finetune create --training-file $FILE_ID 
--model $MODEL_NAME 
--wandb-api-key $WANDB_API_KEY 
--n-epochs 10 
--n-checkpoints 5 
--batch-size 8 
--learning-rate 0.0003
{
    "training_file": "file-aab9997-bca8-4b7e-a720-e820e682a10a",
    "model_output_name": "username/togethercomputer/llama-2-13b-chat",
    "model_output_path": "s3://together/finetune/63e2b89da6382c4d75d5ef22/username/togethercomputer/llama-2-13b-chat",
    "Suffix": "Llama-2-13b 1",
    "model": "togethercomputer/llama-2-13b-chat",
    "n_epochs": 4,
    "batch_size": 128,
    "learning_rate": 1e-06,
    "checkpoint_steps": 2,
    "created_at": 1687982945,
    "updated_at": 1687982945,
    "status": "pending",
    "id": "ft-5bf8990b-841d-4d63-a8a3-5248d73e045f",
    "epochs_completed": 3,
    "events": [
        {
            "object": "fine-tune-event",
            "created_at": 1687982945,
            "message": "Fine tune request created",
            "type": "JOB_PENDING",
        }
    ],
    "queue_depth": 0,
    "wandb_project_name": "Llama-2-13b Fine-tuned 1"
}

Forge the AI frontier. Train on expert-built clusters.

Built by AI researchers for AI innovators, Together GPU Clusters are powered by NVIDIA GB200, H200, and H100 GPUs, along with the Together Kernel Collection — delivering up to 24% faster training operations.

  • Top-Tier NVIDIA GPUs

    NVIDIA's latest GPUs, like GB200, H200, and H100,
for peak AI performance, supporting both training and inference.

  • Accelerated Software Stack

    The Together Kernel Collection includes
custom CUDA kernels, reducing training times and costs with superior throughput.

  • High-Speed Interconnects

    InfiniBand and NVLink ensure fast
communication between GPUs,
eliminating bottlenecks and enabling
rapid processing of large datasets.

  • Highly Scalable & Reliable

    Deploy 16 to 1000+ GPUs across global locations, with 99.9% uptime SLA.

  • Expert AI Advisory Services

    Together AI’s expert team offers
consulting for custom model development
and scalable training best practices.

  • Robust Management Tools

    Slurm and Kubernetes orchestrate
dynamic AI workloads, optimizing training
and inference seamlessly.

Together GPU Clusters

Training-ready clusters – H100, H200, or A100

Reserve your cluster today

THE AI
ACCELERATION
CLOUD

BUILT ON LEADING AI RESEARCH.

Sphere

Innovations

Our research team is behind breakthrough AI models, datasets, and optimizations.

See all research

Cocktail SGD

With Cocktail SGD, we’ve addressed a key hindrance to training generative AI models in a distributed environment: networking overhead. Cocktail SGD is a set of optimizations that reduces network overhead by up to 117x.

Read more

FlashAttention-3

FlashAttention-3 achieves up to 75% GPU utilization on H100s, making AI models up to 2x faster and enabling efficient processing of longer text inputs. It allows for faster training and inference of LLMs, supports lower precision operations for improved efficiency.

Read more

RedPajama

Our RedPajama project enables leading generative AI models to be available as fully open-source. The RedPajama models have been downloaded millions of times, and the RedPajama dataset has been used to create over 500 leading models.

Read more

Sub-quadratic model architectures

In close collaboration with Hazy Research, we’re working on the next core architecture for generative AI models that will provide even faster performance with longer context. Our research published in this area includes Striped Hyena, Monarch Mixer, and FlashConv.

Read more

Customer Stories

See how we support leading teams around the world. Our customers are creating innovative generative AI applications, faster.

Pika creates the next gen text-to-video models on Together GPU Clusters

Nexusflow uses Together GPU Clusters to build cybersecurity models

Arcee builds domain adaptive language models with Together Custom Models

Start
building
yours
here →

Subscribe to newsletter

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.