GPT-5.4 Pro (OpenAI)

GPT-5.4 Pro sits at the top of OpenAI's GPT-5.4 generation as a proprietary, reasoning-oriented system, paired with a same-generation non-Pro variant and distinguished by Azure and OpenAI hosting options. It is positioned as a frontier release that unifies reasoning, coding, and agentic capabilities in a single interface, designed for tasks where sustained deliberation matters more than raw chat fluency. The model is exposed through the Responses API and is also reachable from Codex and Amazon Bedrock, suggesting an architecture intended for long, tool-heavy workflows rather than quick turn-taking.

In benchmark tracking, GPT-5.4 Pro earns a composite score of 61.5 out of 100, ranking 55th of 228 models as of late August 2026, with its strongest category clearly being Agentic work. Independent comparisons highlight it as a strong choice for coding agents, browser research, and computer-use scenarios, while leaving it a few points behind newer reasoning leaders on harder reasoning suites such as ARC-AGI-2 and competitive but not top-ranked on BrowseComp. Practical teams should weigh that agentic strength against the premium pricing and confirm fit with pilot workloads before large-scale adoption.

LLM Gatewayopenai/gpt-5.4-progpt-pro

Open provider page Provider docs

Quick Info

Provider: LLM Gateway
Model key: openai/gpt-5.4-pro
Release date: Mar 5, 2026
Last updated: Mar 5, 2026
Knowledge cutoff: 2025-08-31
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $30.00
Output token cost: $180.00

Limits

Input tokens: 922,000 tokens
Output tokens: 128,000 tokens
Context window: 1,050,000 tokens

Transparent token rates

Compare GPT-5.4 Pro (OpenAI) pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-5.4 Pro (OpenAI)

LLM Gateway

CoverageBenchmark

GPT-5.4 is a frontier language model released by OpenAI on March 5, 2026, unifying reasoning, coding, and agentic capabilities into a single model, with GPT-5.4 Pro offered for maximum performance on complex tasks. According to the page, it is OpenAI's first general-purpose model with native computer-use capabilities ( The model supports text and image inputs plus native computer-use, tool use, web search, file search, function calling, and coding, with a 1,050,000-token context window and 128,000-token max output. Pricing for GPT-5.4 Pro is listed at $30.00 per 1M input tokens and $180.00 per 1M output tokens, with the standard GPT-

LLM Gateway

CoverageBenchmark

BenchmarkList provides aggregated eval results for GPT-5.4 Pro across multiple benchmarks, with pricing listed at $30.00 per 1M input tokens and $180.00 per 1M output tokens. Agentic category shows a median 87th percentile across 6 evals, including MultiChallenge at 69.23 (3rd of 34, 94th percentile), ARC-AGI-2 at 83.3 Among other tracked benchmarks, GPT-5.4 Pro is compared against several models including GPT-5.6 Sol, Claude Fable 5.1, Claude Opus 5, Claude Fable 5, Kimi K3, and Qwen3.8-Flash variants with their respective pricing. The page includes an "Agent Island" benchmark described as a saturation- and contamination-resistant e