← 全球大模型库

GPT-5.4

OpenAI · openai/gpt-5.4

价格 $2.2 起 ~ $5 / 百万 tokens · 共 26 家算力服务商部署

各服务商对比

服务商模型 ID上下文最大输出输入 $/M输出 $/M缓存 $/MAPI 端点

Poe openai/gpt-5.4 1.1M 128K $2.2 $14 $0.22 api.poe.com ↗

Requesty openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 router.requesty.ai ↗

Vercel AI Gateway openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

302.AI gpt-5.4 1.1M 128K $2.5 $15 $0.25 api.302.ai ↗

Vivgrid gpt-5.4 400K 128K $2.5 $15 $0.25 api.vivgrid.com ↗

OpenAI gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

Merge Gateway openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

NEAR AI Cloud openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 cloud-api.near.ai ↗

LLM Gateway gpt-5.4 1.1M 128K $2.5 $15 $0.25 api.llmgateway.io ↗

Abacus gpt-5.4 1.1M 128K $2.5 $15 — routellm.abacus.ai ↗

Cloudflare AI Gateway openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

Azure gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

FreeModel gpt-5.4 1.1M 128K $2.5 $15 $0.25 cc.freemodel.dev ↗

SAP AI Core gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

OpenCode Zen gpt-5.4 1.1M 128K $2.5 $15 $0.25 opencode.ai ↗

Perplexity Agent openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 api.perplexity.ai ↗

Kilo Gateway openai/gpt-5.4 1.1M 128K $2.5 $15 — api.kilo.ai ↗

OpenRouter openai/gpt-5.4 1.1M 128K $2.5 $15 $0.25 openrouter.ai ↗

Azure Cognitive Services gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

AIHubMix gpt-5.4 1.1M 128K $2.5 $15 $0.25 —

NanoGPT openai/gpt-5.4 922K 128K $2.5 $15 $0.25 nano-gpt.com ↗

GitHub Copilot gpt-5.4 400K 128K $2.5 $15 $0.25 api.githubcopilot.com ↗

Cortecs gpt-5.4 1.1M 128K $3 $16.13 $0.25 api.cortecs.ai ↗

ZenMux openai/gpt-5.4 1.1M 128K $3.75 $18.75 — zenmux.ai ↗

OrcaRouter openai/gpt-5.4 1.1M 128K $5 $22.5 $0.25 api.orcarouter.ai ↗

AnyAPI openai/gpt-5.4 1.1M 128K — — — api.anyapi.ai ↗

模型参数

上下文窗口1.1M

最大输出128K

知识截止2025-08-31

发布日期2026-03-05

系列gpt

最后更新2026-03-05

输入模态文本图像PDF

输出模态文本

能力推理工具调用结构化输出文件附件

基准跑分 SWE-Bench Pro 59.1 (resolve rate) · SWE-Atlas Codebase QnA 40.8 (score) · SWE-Atlas Codebase QnA 36.3 (score) · SWE-Atlas Refactoring 44.29 (score) · SWE-Atlas Test Writing 44.36 (score) · SWE-Atlas Test Writing 40 (score) · Artificial Analysis Coding Agent Index 53.6 (average pass@1) · SWE-Atlas Codebase QnA 72.4 (pass@1) · SWE-Bench Pro 18.4 (pass@1) · Terminal-Bench 69.8 (pass@1) · Artificial Analysis Coding Agent Index 52.2 (average pass@1) · SWE-Atlas Codebase QnA 72.9 (pass@1) · SWE-Bench Pro 18.9 (pass@1) · Terminal-Bench 64.7 (pass@1) · Terminal-Bench 75.1 (success rate) · GPQA Diamond 92.8 (accuracy) · Humanity's Last Exam 39.8 (accuracy) · Humanity's Last Exam 52.1 (accuracy) · OSWorld-Verified 75 (success rate) · BrowseComp 82.7 (accuracy) · GDPval 83 (wins or ties) · ARC-AGI-2 73.3 (accuracy) · FrontierMath 47.6 (accuracy) · FrontierMath 27.1 (accuracy) · MMMU Pro 81.2 (accuracy)

OpenAI 的其他模型

GPT-5.1 Codex mini GPT-Image-2 GPT-5.2 Pro GPT-5.5 Pro GPT-4.1 mini GPT-4o GPT-5.4 Pro GPT-5.6 Sol GPT OSS Safeguard 120B GPT-4.1 nano GPT OSS 20B GPT-5.5 Instant GPT-5.2 Chat o3-mini GPT-5.5 o3-pro

数据来自开源项目 models.dev · 价格为美元 / 百万 tokens,以各服务商官方公示为准。