The final major snapshot of the GPT-3.5 generation, released January 2024, fixed a UTF-8 encoding bug affecting non-English function calls and improved format-following accuracy for JSON, YAML, and XML outputs. The gpt-3.5-turbo alias now points permanently to this version, representing the ceiling of what the architecture could achieve after over a year of iterative refinement.
GPT-3.5 Turbo 0125 shares the same 70.0% MMLU (5-shot) and 7.94 MT-Bench baseline as the broader GPT-3.5 Turbo family. As the final snapshot in the series, it improved format-following accuracy and non-English function calling without changing core benchmark performance. Compared to Mixtral 8x7B Instruct (70.6% MMLU, 8.30 MT-Bench) on the sheet, it trails slightly on both measures.
The cost per 1,000 tokens for the model with Telnyx Inference is $0.0010. To illustrate, if an organization were to analyze 1,000,000 customer chats, and each chat consisted of an average of 1,000 tokens, the total cost would be $1,000.
Discover the power and diversity of large language models available with Telnyx. Explore the options below to find the perfect model for your project.
| Organization | Model Name | Tasks | Languages Supported | Context Length | Parameters | Model Tier | License |
|---|---|---|---|---|---|---|---|
| No data available at this time, please try again later. |
Powered by our own GPU infrastructure, select a large language model, add a prompt, and chat away. For unlimited chats, sign up for a free account on our Mission Control Portal here.
Check out our helpful tools to help get you started.
GPT-3.5 Turbo 0125 is not free through the API but is one of OpenAI's most affordable models. It is available through inference providers and may be accessible in ChatGPT's free tier.
Yes, GPT-3.5 Turbo 0125 is the latest snapshot in the 3.5 Turbo series and remains available through the API. OpenAI recommends GPT-4o mini as the successor for new projects.
GPT-3.5 Turbo 0125 improved accuracy on format-following tasks and fixed a bug that caused incomplete UTF-8 sequences. It is a reliable choice for structured output tasks like classification and summarization, though newer models outperform it on reasoning.
GPT-3.5 Turbo 0125 is priced at $0.50 per million input tokens and $1.50 per million output tokens, making it one of the cheapest options in OpenAI's model lineup.
The GPT-3.5 Turbo API is paid with usage-based pricing. Free access is available through ChatGPT with usage limits. For production deployments, hosted inference platforms offer API access with their own pricing.
GPT-4 is significantly more capable than GPT-3.5 Turbo on reasoning, coding, and complex tasks. GPT-3.5 Turbo's advantage is speed and cost, making it better for high-volume, straightforward tasks where maximum quality is not critical.
