OpenAI Reduces GPT-5.6 Turbo Model Pricing, Flagship Unchanged

OpenAI has implemented significant price adjustments across its GPT-5.6 model series. The `gpt-5.6-turbo-256k` model has seen its input token pricing reduced by 80%, moving from $1.00 to $0.20 per million tokens. Its output token price dropped by 85%, from $2.00 to $0.30 per million tokens.

Similarly, the `gpt-5.6-turbo-128k` model's input token price has been cut by 50%, decreasing from $0.50 to $0.25 per million tokens. Output token costs for this model now stand at $0.50 per million tokens, down from $1.00, representing a 50% reduction.

In contrast, the flagship `gpt-5.6-4M` model's pricing remains unchanged. Its input tokens are $5.00 per million, and output tokens are $15.00 per million. This stratified pricing approach suggests a strategic differentiation within OpenAI's model offerings.

The price reductions for the `turbo-256k` and `turbo-128k` models follow a trend observed in the broader AI market. Other major providers, including Anthropic and Google, have also adjusted model pricing, particularly for mid-range and general-purpose models. Anthropic's Claude 3 Haiku, for instance, is priced at $0.25 per million input tokens and $1.25 per million output tokens. Google's Gemini 1.5 Flash offers input tokens at $0.35 per million and output tokens at $1.05 per million.

These price adjustments by OpenAI position its `gpt-5.6-turbo-256k` model as a highly competitive option for cost-sensitive applications. Its new input token price of $0.20 per million undercuts both Claude 3 Haiku and Gemini 1.5 Flash. The output token price of $0.30 per million for `gpt-5.6-turbo-256k` also represents a notable reduction compared to competitors.

The retention of `gpt-5.6-4M`'s pricing suggests OpenAI views this model as a premium offering, distinct from the more commoditised `turbo` variants. The `4M` model offers a significantly larger context window, indicating its utility for highly complex or data-intensive tasks where extended context is critical. This approach allows OpenAI to cater to diverse enterprise needs, from cost-optimised general applications to high-performance, specialised workloads.

Industry analysts note that such pricing strategies reflect increasing competition and the maturing of the large language model market. As model capabilities become more standardised across certain tiers, price becomes a more significant differentiator for providers. OpenAI's move could pressure competitors to re-evaluate their own pricing structures, particularly for models aimed at general business applications.

The `turbo` models are designed for high-throughput and efficiency, making them suitable for tasks such as data processing, content generation, and customer support. The `256k` and `128k` context windows, while substantial, are smaller than the `4M` model's offering. This tiered approach enables OpenAI to capture different segments of the market by offering a range of price-performance options.

This recalibration of prices raises questions about the future trajectory of AI model commoditisation and its impact on the development ecosystem.