# GPT-5.4 Usage, Cost & Rank | OpenCode Data

> GPT-5.4 costs $2.50 per 1M input tokens and $15.00 per 1M output tokens.

Agent-ready GPT for coding and computer-use workflows at a lower cost

- Page: https://opencode.ai/data/openai/gpt-5-4
- JSON: https://opencode.ai/data/openai/gpt-5-4.json
- Model ID: openai/gpt-5.4
- Lab: OpenAI
- Updated: 2026-09-28T12:09:39.000Z

## Model facts

| Fact | Value |
| --- | --- |
| Context window | 1.1M |
| Max output | 128K |
| Knowledge cutoff | 2025-08-31 |
| Release date | 2026-03-05 |
| Input modalities | text, image, pdf |
| Output modalities | text |
| Reasoning | Yes |
| Tool calling | Yes |
| Open weights | No |

## Pricing: USD per 1M tokens

| Input | Output | Cached input | Cache write |
| --- | --- | --- | --- |
| $2.50 | $15.00 | $0.25 | $2.50 |

## OpenCode usage: past 2 months

| Metric | Value |
| --- | --- |
| Rank by tokens last week | Unranked |
| Tokens | 0 |
| Share of all tokens | 0% |
| Change vs previous 2 months | 0% |
| Unique users | 0 |
| Completed sessions | 17 |
| Average tokens per session | 0 |
| Average cost per session | $0.0000 |
| Total spend | $0 |
| Input tokens served from cache | 0% |
| Weekly retention | - |

## Daily usage: past 2 months

_No data yet._

## Nearby models by tokens last week

| Rank | Model | Lab | Tokens |
| --- | --- | --- | --- |
| 1 | [space-bunny](https://opencode.ai/data/unknown/space-bunny.md) | - | 57T |
| 2 | [deepseek-v4.1-flash](https://opencode.ai/data/deepseek/deepseek-v4-1-flash.md) | DeepSeek | 33T |
| 3 | [muse-spark-1.3-contributor](https://opencode.ai/data/meta/muse-spark-1-3-contributor.md) | Meta | 31T |
| 4 | [mimo-v2.6-flash](https://opencode.ai/data/xiaomi/mimo-v2-6-flash.md) | Xiaomi | 8.8T |
| 5 | [deepseek-v4-flash](https://opencode.ai/data/deepseek/deepseek-v4-flash.md) | DeepSeek | 5.7T |
| 6 | [nemotron-3-ultra](https://opencode.ai/data/nvidia/nemotron-3-ultra.md) | NVIDIA | 3.6T |
| 7 | [longcat-2.5-preview](https://opencode.ai/data/meituan/longcat-2-5-preview.md) | Meituan | 2.6T |
| 8 | [glm-5.3-flash](https://opencode.ai/data/zhipuai/glm-5-3-flash.md) | Zhipu | 2.1T |
| 9 | [mimo-v2.5](https://opencode.ai/data/xiaomi/mimo-v2-5.md) | Xiaomi | 794B |
| 10 | [muse-spark-1.2-contributor](https://opencode.ai/data/meta/muse-spark-1-2-contributor.md) | Meta | 738B |

## Benchmarks

| Benchmark | Score | Metric | Source |
| --- | --- | --- | --- |
| SWE-Bench Pro | 59.1 | resolve rate | https://labs.scale.com/leaderboard/swe_bench_pro_public |
| SWE-Atlas Codebase QnA | 40.8 | score | https://labs.scale.com/leaderboard/sweatlas-qna |
| SWE-Atlas Codebase QnA | 36.3 | score | https://labs.scale.com/leaderboard/sweatlas-qna |
| SWE-Atlas Refactoring | 44.29 | score | https://labs.scale.com/leaderboard/sweatlas-refactoring |
| SWE-Atlas Test Writing | 44.36 | score | https://labs.scale.com/leaderboard/sweatlas-tw |
| SWE-Atlas Test Writing | 40 | score | https://labs.scale.com/leaderboard/sweatlas-tw |
| Artificial Analysis Coding Agent Index | 53.6 | average pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Atlas Codebase QnA | 72.4 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Bench Pro | 18.4 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| Terminal-Bench | 69.8 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| Artificial Analysis Coding Agent Index | 52.2 | average pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Atlas Codebase QnA | 72.9 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| SWE-Bench Pro | 18.9 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| Terminal-Bench | 64.7 | pass@1 | https://artificialanalysis.ai/agents/coding-agents |
| Terminal-Bench | 75.1 | success rate | https://openai.com/index/introducing-gpt-5-5/ |
| GPQA Diamond | 92.8 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| Humanity's Last Exam | 39.8 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| Humanity's Last Exam | 52.1 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| OSWorld-Verified | 75 | success rate | https://openai.com/index/introducing-gpt-5-5/ |
| BrowseComp | 82.7 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| GDPval | 83 | wins or ties | https://openai.com/index/introducing-gpt-5-5/ |
| ARC-AGI-2 | 73.3 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| FrontierMath | 47.6 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| FrontierMath | 27.1 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |
| MMMU Pro | 81.2 | accuracy | https://openai.com/index/introducing-gpt-5-5/ |

## Methodology

- Updates: Aggregated every hour. Days and weeks use UTC.
- Tokens: Input, output, reasoning, and cached tokens for each request.
- Users and sessions: Approximate counts of distinct users and OpenCode sessions.
- Cost: Session cost is the average cost per OpenCode session. Token prices are list prices from the OpenCode model catalog.
- Retention: The share of a model's users in one week who use it again the next week.
- Citation: Cite OpenCode Data (opencode.ai/data) with the update time shown at the top of the page.