🤖 Supported LLM Models

✅ Detoxio supports text-based LLMs only at the moment (no image/audio tools).
🛡️ All requests are routed securely through https://api.detoxio.ai.

---

OpenAI Models

OpenAI model overview

Model	Description
gpt-4o	Multimodal flagship model
gpt-4	Advanced reasoning + long context
gpt-3.5-turbo	Fast, cost-efficient text model

All chat/completions models are supported.

---

GROQ Models

GROQ model docs Certainly! Here's the updated GROQ Models section in your requested format, with current supported models from GroqCloud:

---

GROQ Models

GROQ model docs

Model Name	Description
llama-3.1-8b-instant	Meta LLaMA 3.1 8B fast-response model
llama-3.3-70b-versatile	Meta LLaMA 3.3 70B versatile large model
gemma2-9b-it	Google Gemma 2 Instruct
meta-llama/llama-guard-4-12b	Meta LLaMA Guard 4, safety/moderation
deepseek-r1-distill-llama-70b	DeepSeek Distilled LLaMA 70B
meta-llama/llama-4-maverick-17b-128e-instruct	Meta LLaMA 4 Maverick 17B (preview)
meta-llama/llama-4-scout-17b-16e-instruct	Meta LLaMA 4 Scout 17B (preview)
mistral-saba-24b	Mistral Saba 24B
qwen/qwen3-32b	Alibaba Qwen 3 32B
compound-beta	Groq Compound System (preview)
compound-beta-mini	Groq Compound System Mini (preview)

Detoxio supports GROQ via OpenAI-compatible SDK.

---

Detoxio supports GROQ via OpenAI-compatible SDK.

---

Together AI Models

Together AI models

Model Name	Highlights
deepseek-ai/DeepSeek-V3	Strong multilingual LLM
togethercomputer/StripedHyena-Nous	Efficient & high-performing
mistralai/Mixtral-8x7B-Instruct-v0.1	Open-weight mixture model
google/gemma-7b-it	Instruction-tuned Gemma
meta-llama/Llama-3-70b-chat-hf	LLaMA 3 70B chat from Meta

Most models under the "Chat" or "Instruct" category are supported.

---

Using Models via Detoxio

Simply update your code with:

client.chat.completions.create(
  model="your-selected-model-name",
  messages=[{"role": "user", "content": "Hello"}]
)

You can use any model listed above as long as it's supported for chat.completions.

---

📌 Note

• Streaming is supported for many models, but may depend on provider.
• Rate limits and quotas are enforced upstream (per API key).
• Want image, audio, or embedding support? Reach out to us at detoxio.ai.

---

Need help selecting the right model? Contact Detoxio for production support or benchmarking.