Proprietary and open-weight models
Access GPT, Claude, Gemini, Kimi, GLM, DeepSeek, MiMo and more through TextCortex. Choose models from the catalog available to your account.
TextCortex API · Models and routing
Put a number on your API workload.
Estimate TextCortex API token costs and compare models using your expected traffic, input tokens and output tokens.
One API. Every model family.
GPTOpenAI
ClaudeAnthropic
GeminiGoogle
GLM5.3
DeepSeek
KimiK3
MiMo
TextCortex
Why TextCortex
Compare the models your application actually uses.
One API makes it easier to evaluate proprietary and open-weight models using the same workload. Enter the rates for your account, then compare cost with output quality and latency.
One OpenAI-compatible API
Use a familiar client, one TextCortex API key and a common base URL. Change the model identifier to switch between supported models.
EU-hosted model options
Use EU-hosted routes for European inference requirements. Hosting depends on the selected model and route; the wider catalog also includes other deployments.
Cost calculator
Make the token math work for you.
Enter the input and output rates from your quote or chosen model. Use the same currency for both rates.
Estimated monthly token cost
An estimate using your inputs, not a TextCortex price quote. Excludes taxes, cache discounts, retries and additional billable usage. Check the rates for your selected model.
Price input and output separately
Measure how many input and output tokens your application sends and receives. Multiply each quantity by the applicable rate, then divide by the rate unit. Do not assume that a model’s input and output prices are identical.
Use the calculator with the current rates for your model and account. These pages do not publish a fixed rate card or imply unlimited API usage through a workspace subscription.
Model selection changes the bill
A longer response, a reasoning-heavy request or a repeated attempt can consume more tokens than your first successful test. Collect usage across representative requests before forecasting monthly spend.
Compare GPT, Claude and Gemini with Kimi, GLM, DeepSeek and MiMo where each is suitable for the task. A less expensive request only saves money if it produces a usable result.
Build a routing budget
Define a model and a budget for each application feature. This makes a change in traffic or model choice easier to understand.
- Forecast monthly requests for each feature.
- Measure typical input, output and repeated attempts.
- Apply the selected model’s current rates and billing rules.
- Track latency and successful outcomes alongside token cost.
Start with a small measured workload
Get API access, select a model and run a controlled sample. Use the results to set application limits and decide where a different model is worthwhile.
Read the cost-per-task guide for a practical method that includes invalid responses and retries.
Questions
Your next questions, answered.
How is API usage priced?
Use the rates for your selected model and account. Estimate input and output tokens separately and include any applicable cached-token, reasoning-token or other charges. API consumption is distinct from a workspace subscription.
Is TextCortex limited to open-source models?
No. The API provides access to proprietary model families such as GPT, Claude and Gemini alongside open-weight families such as Kimi, GLM, DeepSeek and MiMo. Use GET /models to list the identifiers available to your account.
How does model routing work?
Your application selects an available model in each API request, and TextCortex routes the request through its API. You can choose different models for different tasks. Automatic selection and fallback behavior should only be used where documented for your configuration.
TextCortex AI
One integration. Your choice of models.
Sign up for TextCortex, create an API key and start building with the models your application needs.
Sources and documentation
Product documentation and provider references used for this guide. Reviewed on 5 October 2026.