Together AI
Open-source model inference API + fine-tuning platform, excellent value
What is Together AI
Together AI is an inference and fine-tuning platform focused on open-source models. Offers APIs for dozens of models including Llama, Mixtral, DeepSeek at 5-10x cheaper than OpenAI.
$5 free credits for new users, OpenAI-compatible API format for easy migration. Also supports fine-tuning with your own data.
$5 free credits for new users, OpenAI-compatible API format for easy migration. Also supports fine-tuning with your own data.
Free Tier & Pricing
Free credits: $5 (~5M tokens of Llama 3.3 70B)
Popular model pricing:
- Llama 3.3 70B: $0.88/M tokens
- Mixtral 8x22B: $0.60/M tokens
- DeepSeek V3: $0.90/M tokens
- Qwen 2.5 72B: $0.90/M tokens
Fine-tuning: $3/hr (A100), supports LoRA and full fine-tuning.
Popular model pricing:
- Llama 3.3 70B: $0.88/M tokens
- Mixtral 8x22B: $0.60/M tokens
- DeepSeek V3: $0.90/M tokens
- Qwen 2.5 72B: $0.90/M tokens
Fine-tuning: $3/hr (A100), supports LoRA and full fine-tuning.
Editor's note
Editor's note: If you only need API inference, you may not need a GPU rental. Compare free quota, rate limits, and latency first.
FAQ
Q: How does it compare to Groq?
A: Groq is faster (dedicated hardware), Together AI has more models and supports fine-tuning.
Q: OpenAI API compatible?
A: Yes, just change the base_url to migrate.
Q: Accessible from China?
A: Direct access available with acceptable latency.
A: Groq is faster (dedicated hardware), Together AI has more models and supports fine-tuning.
Q: OpenAI API compatible?
A: Yes, just change the base_url to migrate.
Q: Accessible from China?
A: Direct access available with acceptable latency.