This frozen March 2023 checkpoint was one of the rarest OpenAI models ever offered, restricted behind a separate waitlist even after GPT-4 API access opened broadly in July 2023. At $0.06/$0.12 per thousand tokens, it was the most expensive chat model OpenAI sold, and was made obsolete within months when GPT-4 Turbo delivered 128K context at a lower price.
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.
GPT-4 32K extends the standard GPT-4's 8K context window to 32,768 tokens, allowing it to process longer documents and conversations in a single request. The underlying model architecture and capabilities are identical, with the larger context window being the only difference.
GPT-4 32K pricing on Azure is $60 per million input tokens and $120 per million output tokens, making it significantly more expensive than standard GPT-4. Alternative platforms like Telnyx provide access to GPT-4 variants through their inference infrastructure.
OpenAI is gradually retiring older GPT-4 snapshots as newer models like GPT-4o and GPT-4.1 offer better performance at lower cost. The was one of the earliest GPT-4 releases and has been superseded by multiple iterations.
GPT-4 32k shares the same model weights and 86.4% MMLU (5-shot) score as the standard GPT-4, with the 32k context window being the only architectural difference. On MT-Bench it scored 8.99, the highest of any model at launch. Compared to GPT-3.5 Turbo (70.0% MMLU) on the same sheet, it represents a 16-point improvement in general knowledge.
Running GPT-4 32k through Telnyx Inference costs $60.00 per million input tokens and $120.00 per million output tokens. Processing 100,000 long-document analyses at 10,000 tokens each would cost approximately $60,000 input plus $120,000 output. For most workloads, GPT-4 Turbo ($10/$30 per million tokens) on the same sheet offers comparable quality at one-sixth the cost.
GPT-4 128K refers to GPT-4 Turbo, which expanded the context window from 32K to 128,000 tokens while reducing API pricing. It also added features like JSON mode and improved instruction following that were not available in the original GPT-4 32K.
GPT-5 currently holds the top position across most benchmarks, followed by GPT-4.1 and GPT-4o. The GPT-4 32K 0314 snapshot was considered state-of-the-art at launch in March 2023 but has since been surpassed by multiple generations.
