New GLM-4.7 & MiniMax M2.1 are now live. Try it now
Sign up

33+ private AI models. One sovereign API.

Access the world’s leading open-weight models through a single OpenAI-compatible endpoint. All inference runs on EU-sovereign infrastructure with zero data retention.

Browse Models Open Arena
Supported Models

Access all major AI models through a single API

Tensorx supports leading providers with intelligent routing and automatic failover.
All models run on EU-sovereign infrastructure with zero data retention.

Showing 28 of 28 models

meta-llama/llama-4-maverick

meta-llama 1050K
Input $0.14 / 1M
Output $0.68 / 1M
Functions Tool Choice

deepseek/deepseek-chat-v3-0324

deepseek 163.8K
Input $0.30 / 1M
Output $1.00 / 1M
Functions Tool Choice

deepseek/deepseek-chat-v3.1

deepseek 164K
Input $0.20 / 1M
Output $0.80 / 1M
Functions Tool Choice Reasoning

deepseek/deepseek-v3.2

deepseek 163.8K
Input $0.30 / 1M
Output $0.50 / 1M
Functions Tool Choice Reasoning

meta-llama/llama-3.3-70b-instruct

meta-llama 131K
Input $0.10 / 1M
Output $0.31 / 1M
Functions Tool Choice

minimax/minimax-m2.1

minimax 196.6K
Input $0.30 / 1M
Output $2.40 / 1M
Functions Tool Choice Reasoning Vision

minimax/minimax-m2.5

minimax 65.5K
Input $0.30 / 1M
Output $1.20 / 1M
Functions Tool Choice Reasoning

mistralai/mixtral-8x22b-instruct

mistralai 66K
Input $2.00 / 1M
Output $6.00 / 1M
Functions Tool Choice

moonshotai/kimi-k2.5

moonshotai 262.1K
Input $0.50 / 1M
Output $2.80 / 1M
Functions Tool Choice Vision

moonshotai/Kimi-K2.6

New moonshotai 262.1K
Input $1.00 / 1M
Output $4.00 / 1M
Functions Tool Choice Reasoning Vision

nvidia/nemotron-3-super-120b-a12b

nvidia 262.1K
Input $0.30 / 1M
Output $0.90 / 1M
Functions Tool Choice Reasoning

openai/gpt-oss-120b

openai 131K
Input $0.04 / 1M
Output $0.20 / 1M
Functions Tool Choice Reasoning

openai/gpt-oss-20b

openai 131K
Input $0.03 / 1M
Output $0.14 / 1M
Functions Tool Choice Reasoning

qwen/qwen-2.5-72b-instruct

qwen 33K
Input $0.07 / 1M
Output $0.26 / 1M
Functions Tool Choice

qwen/qwen3-235b-a22b-2507

qwen 131K
Input $0.07 / 1M
Output $0.46 / 1M
Functions Tool Choice

qwen/qwen3.5-122b-a10b

qwen 262.1K
Input $0.50 / 1M
Output $3.50 / 1M
Functions Tool Choice Reasoning

qwen/qwen3.5-9b

qwen 262.1K
Input $0.15 / 1M
Output $0.20 / 1M
Functions Tool Choice Reasoning

qwen/qwen3-coder-30b-a3b-instruct

qwen 262K
Input $0.06 / 1M
Output $0.25 / 1M
Functions Tool Choice

qwen/qwen3-embedding-8b

qwen 32.8K
Input $0.01 / 1M
Output $0.00 / 1M

qwen/qwen3-vl-235b-a22b-instruct

qwen 131K
Input $0.21 / 1M
Output $1.90 / 1M

Systran/faster-whisper-large-v3

systran
Input -
Output -

chatterbox-turbo

unknown
Input $5.00 / 1M
Output -

z-ai/glm-4.6

z-ai 203K
Input $0.40 / 1M
Output $1.75 / 1M
Functions Tool Choice Reasoning

z-ai/glm-4.7

z-ai 202.8K
Input $0.60 / 1M
Output $2.20 / 1M
Functions Tool Choice

z-ai/glm-5

New z-ai 202.8K
Input $1.00 / 1M
Output $3.20 / 1M
Functions Tool Choice Reasoning

z-ai/glm-5.1

New z-ai 202.8K
Input $1.40 / 1M
Output $4.40 / 1M
Functions Tool Choice Reasoning

minimax/minimax-m2

minimax 196.6K
Input $0.25 / 1M
Output $1.00 / 1M
Functions Tool Choice Reasoning

deepseek/deepseek-r1-0528

deepseek 164K
Input $0.66 / 1M
Output $2.60 / 1M
Functions Tool Choice Reasoning
Model Arena

Compare AI Models Side by Side.

Not sure which model to use? The Tensorx Arena lets you run the same prompt across up to four models simultaneously and compare responses in real time — all on EU-sovereign infrastructure with zero data retention.

Real-time side-by-side responses

Select up to 4 models and compare their answers to the same prompt simultaneously.

Ranked by human preference

Model rankings are powered by 50,000+ human preference votes via LMArena.ai.

Task-specific presets

One-click presets for Coding, Finance, Legal, and General reasoning tasks.

Zero data retention

All Arena queries run on EU-sovereign Tensorx infrastructure. Nothing is logged.

Open Model Arena
Usage Example

Drop-in replacement for the OpenAI SDK

Tensorix is fully compatible with the OpenAI API. Change two lines of code — your API key and the base URL — and you are running private, EU-sovereign inference with zero data retention.

  • Compatible with all OpenAI SDK versions
  • Works with LangChain, LlamaIndex, and other frameworks
  • No code changes beyond key + baseURL
  • 33+ models available through the same interface
Get Started Open Arena
inference.py
import OpenAI from 'openai'

 

const client = new OpenAI({

  apiKey: 'YOUR_TENSORIX_API_KEY',

  baseURL: 'https://api.tensorix.ai/v1'

})

 

const response = await client.chat.completions.create({

  model: 'minimax/minimax-m2',

  messages: [{ role: 'user', content: 'Hello!' }]

})

 

console.log(response.choices[0].message.content)