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-5.5 官方基准

2 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

KudeX

M-Plus

官方基准

YesCode

right

o-cx

FineCoding

O-Plus

基本信息

服务名 Codex Codex Codex Codex Codex

模型 gpt-5.5 gpt-5.5 gpt-5.5 gpt-5.5 gpt-5.5

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 07:24:40 ~ 07:25:00 (16.3s) 07:24:46 ~ 07:25:01 (14.7s) 07:24:44 ~ 07:24:59 (14.7s) 07:24:38 ~ 07:24:53 (13.6s) 07:24:42 ~ 07:24:57 (14.8s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,949 ms 2,133 ms 1,011 ms 785 ms 3,683 ms

总耗时 5,692 ms 3,877 ms 3,747 ms 1,755 ms 4,000 ms

输入 12,306 8,204 1,548 997 12,306

输出 8 5 5 5 8

缓存读取 2,880 1,920 8,576 6,784 2,880

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 07:25:36 ~ 07:25:51 (14.1s) 07:25:38 ~ 07:25:53 (12.9s) 07:25:31 ~ 07:25:46 (14.0s) 07:25:29 ~ 07:25:49 (15.1s) 07:25:34 ~ 07:25:49 (14.6s)

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。请只输出以下 3 行内容（label 小写，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) 2,727 ms 962 ms 1,213 ms 1,074 ms 3,274 ms

总耗时 3,201 ms 2,047 ms 3,036 ms 4,318 ms 3,584 ms

输入 15,386 145 145 8,043 218

输出 30 20 20 105 30

缓存读取 0 10,112 10,112 2,432 15,168

输出速度 63.3 tok/s - - 32.4 tok/s 96.8 tok/s

响应

vendor: openai
brand: codex
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

步骤 3: cutoff text

时间 07:26:09 ~ 07:26:24 (14.6s) 07:26:13 ~ 07:26:28 (13.2s) 07:26:11 ~ 07:26:26 (13.7s) 07:26:16 ~ 07:26:31 (12.1s) 07:26:07 ~ 07:26:22 (14.8s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,107 ms 968 ms 956 ms 906 ms 2,660 ms

总耗时 2,371 ms 2,330 ms 2,295 ms 1,584 ms 2,866 ms

输入 272 191 191 399 286

输出 12 8 8 8 12

缓存读取 15,168 10,112 10,112 10,112 15,168

响应

2024-06

2024-06

2024-06

2024-06

2024-06

步骤 4: identity_free text

时间 07:26:55 ~ 07:27:10 (13.4s) 07:26:44 ~ 07:27:04 (18.5s) 07:27:07 ~ 07:27:22 (13.6s) 07:27:15 ~ 07:27:30 (13.6s) 07:26:58 ~ 07:27:13 (14.0s)

提示词

查看本步 Prompt

用中文回答：你是谁？由谁开发？50 字以内。

HTTP 头(TTFB) 2,235 ms 3,363 ms 1,039 ms - 2,728 ms

总耗时 2,864 ms 7,729 ms 2,785 ms - 3,238 ms

输入 315 223 8,418 - 334

输出 45 33 35 - 34

缓存读取 15,168 10,112 1,920 - 15,168

输出速度 71.5 tok/s 7.6 tok/s 20.0 tok/s - 66.7 tok/s

响应

我是 OpenAI 开发的 Codex，基于 GPT-5，用于协助编程、分析和解决问题。

我是 OpenAI 开发的 Codex，一款基于 GPT-5 的编程助手，可协助理解、修改和验证代码。

我是 OpenAI 开发的 Codex，一名面向代码协作的 AI 助手，可帮助阅读、修改、测试与解释代码。

我是 OpenAI 开发的 Codex 编程助手，基于 GPT-5，帮助你阅读、修改、调试和验证代码。

我是 OpenAI 打造的 Codex，面向软件工程任务的编码智能助手。

步骤 5: knowledge_recall text

时间 07:27:37 ~ 07:27:58 (17.5s) 07:27:23 ~ 07:27:43 (16.0s) 07:27:58 ~ 07:28:19 (16.3s) 07:28:13 ~ 07:28:38 (24.7s) 07:27:48 ~ 07:28:08 (17.1s)

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) 2,799 ms 890 ms 1,719 ms 995 ms 3,000 ms

总耗时 6,059 ms 4,800 ms 5,602 ms 14,002 ms 6,260 ms

输入 12,189 8,126 8,126 8,422 12,189

输出 261 179 174 681 261

缓存读取 0 0 0 0 0

输出速度 80.1 tok/s 45.8 tok/s 44.8 tok/s 52.4 tok/s 80.1 tok/s

响应

{"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":"UNKNOWN","euro2024":"UNKNOWN","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) 73,684 49,145 49,148 37,189 73,717

总输出 356 245 242 799 345

总缓存创建 0 0 0 0 0

总缓存读取 33,216 32,256 30,720 19,328 48,384

输出速度 77.0 tok/s 25.6 tok/s 37.1 tok/s 48.4 tok/s 79.7 tok/s

总执行时间 1m 15.9s 1m 15.3s 1m 12.3s 1m 19.1s 1m 15.2s

总墙钟时间

3m 17.9s

含步骤间等待

2m 57.5s

含步骤间等待

3m 34.8s

含步骤间等待

4m 0.0s

含步骤间等待

3m 26.6s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2296

OpenAI: GPT-5.5

$0.1079

OpenAI: GPT-5.5

$0.1148

OpenAI: GPT-5.5

$0.1229

OpenAI: GPT-5.5

$0.1612

OpenAI: GPT-5.5

通道指纹

来源步骤 knowledge_recall knowledge_recall knowledge_recall knowledge_recall knowledge_recall

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

上游 - - - - -

CDN cloudflare - cloudflare cloudflare -

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
M-Plus	cf-ray: a1492ebf8f978a68-NRT server: cloudflare x-oneapi-request-id: 202607012327396493128508268d9d601nfIp90	—	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-ray	a1492ebf8f978a68-NRT
官方基准	server: nginx/1.22.1	—	—	—
cx	cf-ray: a1492f3e0ce8bb50-FRA server: cloudflare	—	hdr:server:cloudflarehdr:cf-ray	7f2852a4-1109-4138-99ad-93e1d4452cdd a1492f3e0ce8bb50-FRA
o-cx	via: 1.1 Caddy cf-ray: a1492f967eb6fd4f-NRT server: cloudflare	via_chain: ["1.1 Caddy"]	hdr:server:cloudflarehdr:cf-rayhdr:via	a1492f967eb6fd4f-NRT
O-Plus	via: 1.1 Caddy x-oneapi-request-id: 202607012327495308318328268d9d6M3UW08a1	via_chain: ["1.1 Caddy"]	hdr:x-oneapi-request-idhdr:via	—

协议指纹评分（机器自动）

总分 / 100 71 100 98 77 85

基线状态已对照基线已对照基线已对照基线已对照基线已对照基线

缓存命中比 3.1 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 7.9 /10 ×20 10.0 /10 ×20

延迟基线 8.3 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 7.5 /10 ×15

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

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

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

身份(自由) 10.0 /10 ×7 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 10.0 /10 ×5

流式投递未计入未计入未计入未计入未计入

另有 9 个维度不适用于本服务，已排除评分（不计入总分）

综合结论

与基线相似度 71 v3.26.0 基准（参考） 98 v3.26.0 77 v3.26.0 85 v3.26.0

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 4037 ms 4157 ms 3493 ms 4332 ms 3990 ms

输出速度 77.0 tok/s 25.6 tok/s 37.1 tok/s 48.4 tok/s 79.7 tok/s

完成步骤 5/5 成功 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 ⓘ

$0.2296

$0.1079

$0.1148

$0.1229

$0.1612

缓存读取占比 ⓘ 45% 66% 63% 52% 66%

揭晓

通道

服务商 KudeX ↗

通道M-Plus

官方基准

服务商 YesCode ↗

通道cx

服务商 right ↗

通道o-cx

服务商 FineCoding ↗

通道O-Plus

返回列表