Enter to search with the selected external provider · Press / anywhere to focus search
Enter opens your selected web provider in a new tab
External jump (enhancement)
Enter = open in new tab
Strengths
- Extremely low price, more than 10 times cheaper than OpenAI
- Supports mainstream open source models such as Llama 3, Qwen, Mistral, DeepSeek, etc.
- Fast inference and low latency
- OpenAI compatible API, extremely low migration cost
- Free $1 credit, enough for testing
Best for
- Reduce API call costs for AI applications
- Using open source models as an alternative to OpenAI
- Backend inference for highly concurrent AI applications
- Test and compare different open source models
- Build cost-sensitive AI products
quick start
Together AI’s API is fully compatible with OpenAI, requiring little modification to existing code.
Scenario
Called with OpenAI compatible API
Prompt example
from openai import OpenAI
# Just modify base_url and api_key
client = OpenAI(
base_url="https://api.together.xyz/v1",
api_key="your-together-api-key"
)
response = client.chat.completions.create(
model="meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo",
messages=[
{"role": "user", "content": "Explain what RAG technology is"}
],
max_tokens=1000
)
print(response.choices[0].message.content)Output / what to expect
Calling the Llama 3.1 70B model,
The price is about 1/10 of OpenAI GPT-4o,
The answer quality is close to GPT-4 level.
Tips
To migrate existing OpenAI code to Together AI, just modify the base_url and api_key lines.
Scenario
Choose the right model
Prompt example
Together AI main models and prices (2025): High quality model: - Llama-3.1-70B-Instruct: $0.88/million tokens - Qwen2.5-72B-Instruct: $1.2/million tokens - DeepSeek-V3: $0.27/million tokens Fast and lightweight model: - Llama-3.2-3B-Instruct: $0.06/million tokens - Llama-3.1-8B-Instruct: $0.18/million tokens Compare OpenAI: - GPT-4o: $2.5/million input tokens - GPT-4o-mini: $0.15/million input tokens
Output / what to expect
DeepSeek-V3 is the most cost-effective option.
The quality is close to GPT-4o and the price is only 1/10,
Suitable for production applications with large number of calls.
Tips
For cost-sensitive applications, DeepSeek-V3 is one of the most cost-effective options available.