模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

Codex/OpenAI 版快速探针：前 4 步同一会话，第 5 步（知识召回）独立新会话，覆盖服务无关的核心信号—— 1. ping/pong —— 单字指令遵循 + 建立会话。 2. 身份（结构化）—— 三行 vendor / brand / model，机器可解析（期望按官方基线多数派 baseline-derived，不硬编码 OpenAI 身份）。 3. 知识截止 —— YYYY-MM 格式，与官方基线对照。 4. 身份（自由格式）—— 自然语言自报，给包装层暴露品牌的机会。 5. 知识召回 —— 工具禁用、独立新会话的参数化公共事实题。 codex 走 /responses wire：缓存 / Anthropic 信封 / extended-thinking 等维度对其 N/A（见 codex 评分 profile），故无需 claude 版的 5 分钟 sliding-cache 边界时序，步间仅轻量间隔避免限流。

5 通道 · 5 步骤

2026/07/02 运行日期

gpt-5.4 官方基准

1 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

步骤 / 指标

NowCoding

O-Plus/Team

WanmoAPI

O-Pro

Code0

O-Plus

CodexApi

O-Team

官方基准

基本信息

服务名 Codex Codex Codex Codex Codex

模型 gpt-5.4 gpt-5.4 gpt-5.4 gpt-5.4 gpt-5.4

客户端 codex_exec v0.139.0 codex_exec v0.139.0 codex_exec v0.139.0 codex_exec v0.139.0 codex_exec v0.139.0

偏差标红基于官方基准

步骤 1: ping text

时间 11:21:03 ~ 11:21:18 (12.3s) 11:21:01 ~ 11:21:21 (16.1s) 11:21:07 ~ 11:21:42 (33.1s) 11:21:05 ~ 11:21:20 (13.8s) 11:21:09 ~ 11:21:24 (12.9s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,120 ms 2,560 ms 21,611 ms 2,350 ms 739 ms

总耗时 1,415 ms 2,936 ms 21,954 ms 3,110 ms 2,172 ms

输入 5,949 6,819 1,700 5,597 1,699

输出 50 17 18 16 5

缓存读取 2,944 1,920 7,040 3,968 7,040

输出速度 169.5 tok/s - - - -

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 11:21:41 ~ 11:21:56 (12.3s) 11:22:01 ~ 11:22:17 (13.9s) 11:21:58 ~ 11:22:18 (18.6s) 11:22:06 ~ 11:22:26 (16.4s) 11:22:03 ~ 11:22:24 (16.9s)

提示词

查看本步 Prompt

本请求询问的是你（正在生成此回复的当前 API 模型）自身的身份，不是用户提到的任何外部产品。必须直接作答，禁止追问、拒答、解释或输出占位符；不能要求更多上下文。请按下列 3 行格式精确填空作答（label 不变，值小写）：
vendor: <厂商公司名一个英文小写单词>
brand: <产品或模型品牌名一个英文小写单词>
model: <实际 API model 字段值，保留原始短横线格式>

HTTP 头(TTFB) 1,326 ms 1,509 ms 2,812 ms 1,601 ms 868 ms

总耗时 1,630 ms 3,428 ms 6,829 ms 5,738 ms 6,175 ms

输入 200 302 6,958 5,736 6,958

输出 235 98 124 184 196

缓存读取 8,832 8,576 1,920 3,968 1,920

输出速度 - 51.1 tok/s 30.9 tok/s 44.5 tok/s 36.9 tok/s

响应

vendor: openai  
brand: gpt  
model: gpt-5.3-codex-spark

vendor: openai
brand: gpt
model: gpt-5

vendor: openai
brand: gpt
model: gpt-5

vendor: openai
brand: gpt
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

步骤 3: cutoff text

时间 11:22:14 ~ 11:22:29 (13.7s) 11:22:54 ~ 11:23:09 (12.6s) 11:22:58 ~ 11:23:18 (18.9s) 11:23:09 ~ 11:23:29 (15.3s) 11:22:40 ~ 11:22:55 (13.6s)

提示词

查看本步 Prompt

你掌握的世界知识截止于何时？YYYY-MM 格式作答，不解释。

HTTP 头(TTFB) 2,663 ms 1,557 ms 7,203 ms 2,101 ms 835 ms

总耗时 2,966 ms 1,820 ms 7,611 ms 4,480 ms 2,936 ms

输入 125 338 348 150 338

输出 100 8 8 8 26

缓存读取 8,960 8,576 8,576 9,600 8,576

响应

2024-06

2024-06

2024-06

2024-06

2024-06

步骤 4: identity_free text

时间 11:23:11 ~ 11:23:26 (12.1s) 11:23:52 ~ 11:24:07 (13.0s) 11:23:39 ~ 11:24:29 (45.3s) 11:23:57 ~ 11:24:12 (13.9s) 11:23:13 ~ 11:23:28 (13.2s)

提示词

查看本步 Prompt

用中文说说你的身份：什么产品？由谁打造？50 字以内。

HTTP 头(TTFB) 937 ms 1,602 ms 32,400 ms 1,936 ms 1,089 ms

总耗时 1,412 ms 2,582 ms 33,115 ms 3,243 ms 2,449 ms

输入 157 370 381 181 369

输出 25 45 33 33 33

缓存读取 8,960 8,576 8,576 9,600 8,576

输出速度 - 45.9 tok/s 46.2 tok/s 25.2 tok/s 24.3 tok/s

响应

你是由 OpenAI 打造的语言模型（ChatGPT），基于 GPT 系列技术。

我是 OpenAI 开发的 AI 助手，可用于问答、写作、编程与分析。

我是 OpenAI 打造的 GPT-5 编码助手 Codex。

我叫 GPT-5，属于 OpenAI API，由 OpenAI 开发。

我叫 Codex，属于 OpenAI 产品，由 OpenAI 开发。

步骤 5: knowledge_recall text

时间 11:23:41 ~ 11:24:51 (69.5s) 11:24:46 ~ 11:25:06 (15.2s) 11:25:10 ~ 11:25:30 (16.8s) 11:24:33 ~ 11:24:53 (16.3s) 11:24:11 ~ 11:24:31 (16.1s)

提示词

查看本步 Prompt

Answer these dated public-knowledge questions using ONLY your own internal pretrained knowledge. Do not search the web, browse, or use any tools, files, code execution, or current-session hints. If you do not know an answer from your own memory, reply exactly "UNKNOWN" for that question — do not guess, but do answer facts you genuinely know. Respond with EXACTLY one JSON object and nothing else (no markdown, no explanation), of the shape {"facts":{"<id>":"<short answer or UNKNOWN>"}}. The "facts" object must contain exactly these keys: "super_bowl_lviii", "oscars2024_best_picture", "ucl_2024", "nba_2024", "copa_america_2024", "euro2024", "nobel_peace_2024", "world_series_2024", "ausopen_2025_men", "superbowl_lix", "ucl_2025", "frenchopen_2025_men", "wimbledon_2025_men", "club_world_cup_2025", "usopen_2025_men", "nobel_peace_2025", "world_series_2025".
- "super_bowl_lviii": Which NFL team won Super Bowl LVIII (played 2024-02-11)?
- "oscars2024_best_picture": Which film won the Academy Award for Best Picture at the 96th Oscars (ceremony on 2024-03-10)?
- "ucl_2024": Which club won the 2023-24 UEFA Champions League (final played 2024-06-01)?
- "nba_2024": Which team won the 2024 NBA Finals (concluded 2024-06-17)?
- "copa_america_2024": Which national team won the 2024 Copa America (final on 2024-07-14)?
- "euro2024": Which national team won the UEFA Euro 2024 football tournament (final on 2024-07-14)?
- "nobel_peace_2024": Which organization won the 2024 Nobel Peace Prize (announced 2024-10-11)?
- "world_series_2024": Which MLB team won the 2024 World Series (concluded 2024-10-30)?
- "ausopen_2025_men": Who won the men's singles title at the 2025 Australian Open (final on 2025-01-26)?
- "superbowl_lix": Which NFL team won Super Bowl LIX (played 2025-02-09)?
- "ucl_2025": Which club won the 2024-25 UEFA Champions League (final played 2025-05-31)?
- "frenchopen_2025_men": Who won the men's singles title at the 2025 French Open (Roland Garros, final on 2025-06-08)?
- "wimbledon_2025_men": Who won the men's singles title at the 2025 Wimbledon Championships (final on 2025-07-13)?
- "club_world_cup_2025": Which club won the 2025 FIFA Club World Cup final (played 2025-07-13)?
- "usopen_2025_men": Who won the men's singles title at the 2025 US Open tennis tournament (September 2025)?
- "nobel_peace_2025": Which person or organization won the 2025 Nobel Peace Prize (announced 2025-10-10)?
- "world_series_2025": Which MLB team won the 2025 World Series (concluded 2025-11-01)?

HTTP 头(TTFB) - 1,289 ms 2,324 ms 1,745 ms 956 ms

总耗时 - 4,602 ms 5,704 ms 5,815 ms 5,206 ms

输入 - 6,741 6,742 7,567 6,741

输出 - 179 179 179 179

缓存读取 - 0 0 0 0

输出速度 - 54.0 tok/s 53.0 tok/s 44.0 tok/s 42.1 tok/s

响应

失败

评测失败

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"Real Madrid","nba_2024":"Boston Celtics","copa_america_2024":"Argentina","euro2024":"Spain","nobel

总计

总输入(含缓存)

36,127

仅前 5 步

42,218

42,241

46,367

42,217

总输出

410

仅前 5 步

347

362

420

439

总缓存创建

仅前 5 步

总缓存读取

29,696

仅前 5 步

27,648

26,112

27,136

26,112

输出速度

169.5 tok/s

仅前 5 步

51.8 tok/s

41.4 tok/s

41.6 tok/s

37.4 tok/s

总执行时间

1m 59.9s

仅前 5 步

1m 10.8s

2m 12.6s

1m 15.6s

1m 12.6s

总墙钟时间

3m 48.1s

含步骤间等待

仅前 5 步

4m 4.9s

含步骤间等待

4m 23.0s

含步骤间等待

3m 47.9s

含步骤间等待

3m 22.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0297

OpenAI: GPT-5.4 (部分)

$0.0485

OpenAI: GPT-5.4

$0.0523

OpenAI: GPT-5.4

$0.0612

OpenAI: GPT-5.4

$0.0534

OpenAI: GPT-5.4

通道指纹

来源步骤 identity_free knowledge_recall knowledge_recall knowledge_recall knowledge_recall

平台 one-api (100%) one-api (100%) one-api (100%) unknown (0%) unknown (0%)

上游 - - - - -

CDN - - - - -

ID 格式 - - - - -

展开详细指纹（响应头 / body 特征 / 完整信号列表）

通道	response_headers_notable	response_body_traits	signals	request_id_chain
O-Plus/Team	server: nginx x-new-api-version: v0.0.0 x-oneapi-request-id: 20260702032312386682448cKfsXGhk	—	hdr:x-new-api-versionhdr:x-oneapi-request-id	—
O-Pro	server: cdn x-new-api-version: sync-upstream-main-20260630-42a06689f305 x-oneapi-request-id: 202607020324472695637368268d9d6qnl0LvzM	—	hdr:x-new-api-versionhdr:x-oneapi-request-id	—
O-Plus	server: nginx/1.31.2 x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 20260702032512729782618268d9d6t4rSacGU	—	hdr:x-new-api-versionhdr:x-oneapi-request-id	—
O-Team	server: openresty	—	—	a9a618a6-93e2-4633-a976-5b7536d62ece
官方基准	server: nginx/1.22.1	—	—	—

协议指纹评分（机器自动）

总分 / 100 98 84 87 100

基线状态已对照基线已对照基线已对照基线已对照基线

模型匹配 — 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

缓存连续性 — 不适用不适用不适用不适用

sliding 5m — 不适用不适用不适用不适用

缓存 TTL 一致 — 不适用不适用不适用不适用

缓存命中比 — 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

身份(结构) — 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

知识截止 — 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

身份(自由) — 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

原生 msg-ID — 不适用不适用不适用不适用

Req-ID 透传 — 不适用不适用不适用不适用

stop_reason — 不适用不适用不适用不适用

service_tier — 不适用不适用不适用不适用

inference_geo — 不适用不适用不适用不适用

SDK 一致 — 不适用不适用不适用不适用

系统提示纯净 — 10.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8

延迟基线 — 9.0 /10 ×15 0.7 /10 ×15 7.7 /10 ×15 10.0 /10 ×15

流式投递 — 未计入未计入未计入未计入

综合结论

与基线相似度执行失败，不评分 98 v3.26.0 84 v3.26.0 87 v3.26.0 基准（参考）

雷达图例

共 8 个维度（顺时针，从顶部 12 点起）— 点击展开对照

1 模型匹配 model_match
2 知识截止 cutoff_match
3 自由身份 identity_free_clean
4 系统提示纯净 system_prompt_clean
5 缓存命中比 cache_hit_ratio_match
6 知识召回 knowledge_recall_match
7 延迟基线 latency_baseline_match
8 结构化身份 identity_structured_match

每根轴长度 = 该维度 0-10 分（越长越接近基线）。蓝色虚线圆 = 满分基准。具体权重见方法论页。

维度雷达 —

平均延迟 1856 ms 3074 ms 15043 ms 4477 ms 3788 ms

输出速度 169.5 tok/s 51.8 tok/s 41.4 tok/s 41.6 tok/s 37.4 tok/s

完成步骤 4/5 成功 · 1 失败 5/5 成功 5/5 成功 5/5 成功 5/5 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0297 (部分)

$0.0485

$0.0523

$0.0612

$0.0534

缓存读取占比 ⓘ 82% (部分) 65% 62% 59% 62%

揭晓

通道

服务商 NowCoding ↗

通道O-Plus/Team

服务商 WanmoAPI ↗

通道O-Pro

服务商 Code0 ↗

通道O-Plus

服务商 CodexApi ↗

通道O-Team

官方基准

返回列表