Open Source & Other Providers

Creor supports a wide range of inference providers that host open-source models, specialized models, and fast-inference hardware. Each provider has different strengths -- from raw speed to specialized capabilities.

Provider Comparison

Provider	Key Models	Auth Env Var	Best For
Groq	Llama 3.x, Mixtral, Gemma	GROQ_API_KEY	Ultra-fast inference (LPU hardware)
Together AI	Llama 3.x, Qwen 2.5, DeepSeek, Mixtral	TOGETHER_AI_API_KEY	Wide open-source model selection
DeepInfra	Llama 3.x, Mistral, Qwen	DEEPINFRA_API_KEY	Cost-effective open models
Cerebras	Llama 3.x	CEREBRAS_API_KEY	Wafer-scale inference speed
Mistral	Mistral Large, Codestral, Ministral	MISTRAL_API_KEY	European AI, code generation
Cohere	Command R, Command R+	COHERE_API_KEY	RAG and enterprise search
Perplexity	Sonar Pro, Sonar	PERPLEXITY_API_KEY	Web-grounded, up-to-date answers
xAI	Grok 3, Grok 3 Mini	XAI_API_KEY	Reasoning, real-time knowledge
Vercel	v0 models	VERCEL_API_KEY	Vercel platform integration

Tip

All of these providers follow the same setup pattern: get an API key, set it as an environment variable or in the Settings UI, and reference models with the provider/model-id format.

Groq

Groq runs inference on custom LPU (Language Processing Unit) hardware, delivering some of the fastest token generation speeds available. Ideal for rapid iteration and tasks where latency matters more than model size.

Setup

Sign up at console.groq.com.
Create an API key from the dashboard.
Set the environment variable or add the key in Creor Settings.

export GROQ_API_KEY="gsk_your-key-here"

Configuration

{

"model": "groq/llama-3.3-70b-versatile"

}

Popular Models

Model	Model ID	Parameters
Llama 3.3 70B	llama-3.3-70b-versatile	70B
Llama 3.1 8B	llama-3.1-8b-instant	8B
Mixtral 8x7B	mixtral-8x7b-32768	46.7B (MoE)
Gemma 2 9B	gemma2-9b-it	9B

Together AI

Together AI hosts the widest selection of open-source models, from Llama and Qwen to DeepSeek and Mixtral. It offers both serverless and dedicated inference options.

Setup

Sign up at api.together.xyz.
Create an API key from your account dashboard.
Set the environment variable or add the key in Creor Settings.

export TOGETHER_AI_API_KEY="your-key-here"

Configuration

{

"model": "togetherai/meta-llama/Llama-3.3-70B-Instruct-Turbo"

}

Popular Models

Model	Model ID	Parameters
Llama 3.3 70B Turbo	meta-llama/Llama-3.3-70B-Instruct-Turbo	70B
Qwen 2.5 72B	Qwen/Qwen2.5-72B-Instruct-Turbo	72B
DeepSeek V3	deepseek-ai/DeepSeek-V3	671B (MoE)
Mixtral 8x22B	mistralai/Mixtral-8x22B-Instruct-v0.1	141B (MoE)

DeepInfra

DeepInfra provides cost-effective inference for popular open-source models with competitive pricing and low latency.

Setup

Sign up at deepinfra.com.
Get your API key from the dashboard.
Set the environment variable or add the key in Creor Settings.

export DEEPINFRA_API_KEY="your-key-here"

Configuration

{

"model": "deepinfra/meta-llama/Llama-3.3-70B-Instruct"

}

Cerebras

Cerebras uses wafer-scale engine (WSE) chips to deliver extremely fast inference. Currently supports Llama models with industry-leading tokens-per-second throughput.

Setup

Sign up at cloud.cerebras.ai.
Create an API key from the console.
Set the environment variable or add the key in Creor Settings.

export CEREBRAS_API_KEY="your-key-here"

Configuration

{

"model": "cerebras/llama-3.3-70b"

}

Mistral

Mistral is a European AI company offering models optimized for code generation, multilingual tasks, and efficient inference. Their Codestral model is purpose-built for coding.

Setup

Sign up at console.mistral.ai.
Create an API key from the dashboard.
Set the environment variable or add the key in Creor Settings.

export MISTRAL_API_KEY="your-key-here"

Configuration

{

"model": "mistral/mistral-large-latest"

}

Popular Models

Model	Model ID	Best For
Mistral Large	mistral-large-latest	Complex reasoning, multilingual
Codestral	codestral-latest	Code generation and completion
Ministral 8B	ministral-8b-latest	Fast, lightweight tasks

Cohere

Cohere specializes in enterprise AI with models optimized for retrieval-augmented generation (RAG) and search. Their Command R models excel at grounded, factual responses.

Setup

Sign up at dashboard.cohere.com.
Create an API key from the API keys page.
Set the environment variable or add the key in Creor Settings.

export COHERE_API_KEY="your-key-here"

Configuration

{

"model": "cohere/command-r-plus"

}

Perplexity

Perplexity's Sonar models are grounded in real-time web search results, making them excellent for questions that require up-to-date information about libraries, APIs, or recent changes.

Setup

Sign up at perplexity.ai and access the API section.
Create an API key.
Set the environment variable or add the key in Creor Settings.

export PERPLEXITY_API_KEY="pplx-your-key-here"

Configuration

{

"model": "perplexity/sonar-pro"

}

xAI (Grok)

xAI's Grok models combine strong reasoning with access to real-time knowledge. Grok 3 is competitive with frontier models on coding and reasoning benchmarks.

Setup

Sign up at console.x.ai.
Create an API key from the dashboard.
Set the environment variable or add the key in Creor Settings.

export XAI_API_KEY="xai-your-key-here"

Configuration

{

"model": "xai/grok-3"

}

Available Models

Model	Model ID	Best For
Grok 3	grok-3	Complex reasoning, full capability
Grok 3 Mini	grok-3-mini	Fast reasoning, lower cost

Vercel

Vercel provides model access through the Vercel AI platform. This is useful for teams already using the Vercel ecosystem.

Setup

Go to vercel.com and sign in.
Navigate to your account settings and find the API tokens section.
Create a token with the appropriate permissions.

export VERCEL_API_KEY="your-vercel-token"

Configuration

{

"model": "vercel/v0-1.0-md"

}

OpenRouter

Access 100+ models through a single API key.

BYOK

Use any OpenAI-compatible endpoint.