GPT-5.4 mini (high)

GPT-5.2 Codex (high) is one of the leading models in intelligence and is well-priced compared to similar models. This model supports text and image input, while outputting both text and images. It has a context window of 400k tokens and knowledge updated to August 2025.

GPT-5.3 Codex (high) is one of the leading models in intelligence, however it is quite expensive compared to models in the same price segment. It stands out with fast processing speed, though it tends to be verbose. This model supports text and image input, outputs text, and has a 400k token context window with knowledge updated to August 2025.

GPT-5.4 (xhigh) is one of the leading models in intelligence, capable of processing text and image input and generating text output. It has a large context window and knowledge updated to August 2025. This model is a proprietary model from OpenAI.

GPT-5.4 nano (xhigh) is a leading intelligence model notable for its fast processing speed. This model supports both text and image input, with the ability to generate text output. Despite its powerful capabilities, it is quite expensive and tends to be verbose in its responses. It has a large context window of 400,000 tokens.

GPT-5.5 (high) is a proprietary OpenAI model, ranked highly for intelligence but with a relatively high cost. This model supports text and image input, while generating text output. It stands out for its reasoning capabilities, including extended thinking or chain-of-thought reasoning, to effectively solve complex problems.

GPT-5.5 (low) is a new model from OpenAI, notable for its intelligence and above-average performance speed. The model is highly compact and supports both text and image inputs, while generating text-based outputs though it comes at a steep price of $5 per 1M input tokens (cached: $0.5) and $30 per 1M output tokens.