AI 模型速度排行榜
按实测输出吞吐(tokens/秒)与平均延迟排名。
- OpenAI dall-e-3 — 吞吐: 2000.0 tps, 延迟: 4234.00s
- OpenAI gpt-5-nano — 吞吐: 385.8 tps, 延迟: 11323.82s
- Google Gemini 3.5 Flash — 吞吐: 336.7 tps, 延迟: 10444.89s
- Google Gemini 3.1 Flash-Lite — 吞吐: 330.0 tps, 延迟: 1333.59s
- xAI Grok 4.20 Multi-Agent — 吞吐: 326.0 tps, 延迟: 0.00s
- Google Gemini 3.6 Flash — 吞吐: 313.6 tps, 延迟: 5821.69s
- Google Gemini Omni Flash Preview — 吞吐: 306.6 tps, 延迟: 99229.00s
- OpenAI GPT-5.4 Mini — 吞吐: 247.9 tps, 延迟: 1147.56s
- OpenAI gpt-5-chat-latest — 吞吐: 218.4 tps, 延迟: 6859.40s
- OpenAI GPT-5 — 吞吐: 207.4 tps, 延迟: 28250.60s
- OpenAI o3-mini — 吞吐: 194.2 tps, 延迟: 5194.60s
- OpenAI GPT-5.6 Terra — 吞吐: 187.9 tps, 延迟: 7493.61s
- Google gemini-2.5-flash — 吞吐: 179.6 tps, 延迟: 23565.75s
- OpenAI chatgpt-image-latest — 吞吐: 163.0 tps, 延迟: 25489.00s
- Google Gemini 3.1 Flash-Lite Preview — 吞吐: 159.8 tps, 延迟: 1525.44s
- OpenAI gpt-4o-all — 吞吐: 150.1 tps, 延迟: 2745.71s
- xAI Grok 4.20 (Non-Reasoning) — 吞吐: 148.9 tps, 延迟: 1179.50s
- xAI grok-4-1-fast-reasoning — 吞吐: 147.8 tps, 延迟: 0.00s
- xAI Grok 4.3 — 吞吐: 144.6 tps, 延迟: 650.00s
- Google Nano Banana (Gemini 2.5 Flash Image 🍌) — 吞吐: 143.8 tps, 延迟: 6076.50s
- Google Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image 🍌) — 吞吐: 141.4 tps, 延迟: 10441.72s
- OpenAI GPT Image 2 Develop — 吞吐: 139.7 tps, 延迟: 28024.64s
- OpenAI gpt-5.1-codex-mini — 吞吐: 134.8 tps, 延迟: 0.00s
- Google Gemini 3.1 Pro Preview — 吞吐: 133.9 tps, 延迟: 28196.79s
- OpenAI gpt-image-1-mini — 吞吐: 133.7 tps, 延迟: 46435.50s
- xAI grok-3 — 吞吐: 129.0 tps, 延迟: 0.00s
- OpenAI GPT-5.5 — 吞吐: 126.4 tps, 延迟: 4220.06s
- Anthropic Claude Haiku 4.5 — 吞吐: 122.9 tps, 延迟: 1627.51s
- xAI grok-3-mini — 吞吐: 122.7 tps, 延迟: 0.00s
- OpenAI GPT Image 1.5 — 吞吐: 122.7 tps, 延迟: 53857.50s
- xAI grok-4-0709 — 吞吐: 121.9 tps, 延迟: 851.25s
- Google Gemini 3 Flash Preview — 吞吐: 121.1 tps, 延迟: 6231.13s
- OpenAI gpt-oss-120b — 吞吐: 119.6 tps, 延迟: 0.00s
- OpenAI gpt-4-gizmo-* — 吞吐: 117.4 tps, 延迟: 7668.17s
- DeepSeek DeepSeek-V4-Flash 0731 — 吞吐: 115.2 tps, 延迟: 8390.11s
- DeepSeek DeepSeek-V4-Pro-0813 — 吞吐: 91.4 tps, 延迟: 912.20s
- OpenAI GPT-5.4 — 吞吐: 90.4 tps, 延迟: 2591.28s
- OpenAI gpt-4o-search-preview — 吞吐: 86.8 tps, 延迟: 0.00s
- Google Nano Banana 2 (Gemini 3.1 Flash Image 🍌) — 吞吐: 86.5 tps, 延迟: 21850.33s
- MiniMax MiniMax M2.5 — 吞吐: 85.9 tps, 延迟: 0.00s
- Anthropic Claude Opus 5 — 吞吐: 83.0 tps, 延迟: 12175.67s
- Anthropic Claude Sonnet 5 — 吞吐: 82.8 tps, 延迟: 4905.39s
- OpenAI GPT-5.6 Sol — 吞吐: 82.8 tps, 延迟: 8253.86s
- Alibaba qwen-plus — 吞吐: 81.8 tps, 延迟: 1735.57s
- OpenAI gpt-4.1 — 吞吐: 79.5 tps, 延迟: 2632.00s
- OpenAI GPT-5.1 — 吞吐: 79.5 tps, 延迟: 976.00s
- MiniMax MiniMax M3 — 吞吐: 77.6 tps, 延迟: 1440.80s
- OpenAI gpt-4o-transcribe-diarize — 吞吐: 76.7 tps, 延迟: 0.00s
- xAI grok-4-1-fast-non-reasoning — 吞吐: 76.4 tps, 延迟: 0.00s
- xAI grok-4-fast-reasoning — 吞吐: 74.6 tps, 延迟: 0.00s
- Google Gemini 3.5 Flash-Lite — 吞吐: 73.0 tps, 延迟: 3230.89s
- Anthropic Claude Fable 5 — 吞吐: 71.9 tps, 延迟: 6259.84s
- Anthropic Claude Opus 4.7 — 吞吐: 69.6 tps, 延迟: 2850.62s
- OpenAI GPT-5.6 Luna — 吞吐: 69.6 tps, 延迟: 2764.51s
- Google Gemini 3 Pro Preview — 吞吐: 64.1 tps, 延迟: 0.00s
- DeepSeek DeepSeek-V4-Pro — 吞吐: 63.9 tps, 延迟: 0.00s
- OpenAI gpt-5-mini — 吞吐: 62.9 tps, 延迟: 4125.00s
- Anthropic Claude Sonnet 4.5 — 吞吐: 60.6 tps, 延迟: 1875.09s
- xAI grok-4-fast-non-reasoning — 吞吐: 59.1 tps, 延迟: 0.00s
- xAI Grok Build 0.1 — 吞吐: 58.7 tps, 延迟: 0.00s
- Google Gemini 3.1 Pro Preview Customtools — 吞吐: 58.6 tps, 延迟: 3729.00s
- DeepSeek deepseek-v3.1 — 吞吐: 56.9 tps, 延迟: 3098.00s
- Anthropic Claude Opus 4.8 — 吞吐: 55.0 tps, 延迟: 13817.76s
- Anthropic Claude Sonnet 4.6 — 吞吐: 54.8 tps, 延迟: 2588.24s
- OpenAI gpt-4o-mini — 吞吐: 51.5 tps, 延迟: 1064.25s
- xAI Grok 4.6 — 吞吐: 51.5 tps, 延迟: 1722.42s
- OpenAI gpt-4o-mini-transcribe — 吞吐: 50.6 tps, 延迟: 0.00s
- OpenAI GPT-5.2 — 吞吐: 50.0 tps, 延迟: 1530.33s
- Z.ai GLM 5.2 — 吞吐: 49.4 tps, 延迟: 896.50s
- xAI Grok 4.20 — 吞吐: 46.9 tps, 延迟: 3046.00s
- Anthropic Claude Opus 4.6 — 吞吐: 45.1 tps, 延迟: 5165.82s
- xAI Grok 4.5 — 吞吐: 43.7 tps, 延迟: 6430.33s
- ByteDance Doubao-Seed-2.0-pro — 吞吐: 37.9 tps, 延迟: 0.00s
- OpenAI gpt-3.5-turbo-16k — 吞吐: 34.0 tps, 延迟: 0.00s
- OpenAI gpt-4.1-mini — 吞吐: 33.4 tps, 延迟: 3542.43s
- OpenAI GPT Image 2 — 吞吐: 32.5 tps, 延迟: 37750.51s
- OpenAI GPT Image 1.0 — 吞吐: 31.4 tps, 延迟: 27011.31s
- OpenAI gpt-4o-2024-11-20 — 吞吐: 30.3 tps, 延迟: 0.00s
- Anthropic Claude Opus 4.5 — 吞吐: 29.9 tps, 延迟: 2179.00s
- OpenAI o4-mini — 吞吐: 29.5 tps, 延迟: 0.00s
- OpenAI gpt-4o-image — 吞吐: 28.4 tps, 延迟: 31274.62s
- Moonshot Kimi K3 — 吞吐: 27.7 tps, 延迟: 31929.00s
- Alibaba qwen-max — 吞吐: 27.0 tps, 延迟: 657.00s
- OpenAI GPT-5.4 Nano — 吞吐: 23.6 tps, 延迟: 1088.00s
- ByteDance Doubao-Seedream-5.0-pro — 吞吐: 23.5 tps, 延迟: 174643.50s
- Google Gemini 3.7 Flash — 吞吐: 22.4 tps, 延迟: 0.00s
- Xiaomi MiMo-V2.5 Pro — 吞吐: 21.6 tps, 延迟: 0.00s
- Google Nano Banana Pro (Gemini 3 Pro Image 🍌) — 吞吐: 20.3 tps, 延迟: 50710.45s
- Alibaba qwen3-vl-flash — 吞吐: 20.0 tps, 延迟: 0.00s
- Xiaomi MiMo-V2.5 — 吞吐: 17.2 tps, 延迟: 4275.00s
- OpenAI gpt-3.5-turbo-1106 — 吞吐: 17.0 tps, 延迟: 0.00s
- Yi yi-lightning — 吞吐: 16.8 tps, 延迟: 0.00s
- Google gemini-2.5-pro — 吞吐: 15.3 tps, 延迟: 0.00s
- OpenAI o1 — 吞吐: 13.7 tps, 延迟: 0.00s
- OpenAI gpt-5.1-codex — 吞吐: 13.6 tps, 延迟: 0.00s
- OpenAI gpt-4o-transcribe — 吞吐: 13.4 tps, 延迟: 0.00s
- OpenAI GPT-5.3 Codex — 吞吐: 11.4 tps, 延迟: 0.00s
- Moonshot Kimi K2.6 — 吞吐: 11.2 tps, 延迟: 1857.75s
- OpenAI GPT-5.5 Pro — 吞吐: 9.9 tps, 延迟: 359.00s
- OpenAI gpt-4-turbo-preview — 吞吐: 8.9 tps, 延迟: 0.00s