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

1 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

PackyCode

cx-Zeus

LyraAPI

O-Plus

ModelFlare

O-Pro

CCAPI

O-Api

官方基准

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 20:02:24 ~ 20:02:44 (17.8s) 20:02:28 ~ 20:02:48 (15.4s) 20:02:26 ~ 20:02:41 (13.3s) 20:02:22 ~ 20:02:42 (18.8s) 20:02:30 ~ 20:02:45 (12.4s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,654 ms 3,388 ms 1,629 ms 2,914 ms 991 ms

总耗时 7,015 ms 4,581 ms 2,291 ms 7,724 ms 1,739 ms

输入 7,723 1,548 1,548 1,345 1,548

输出 5 5 5 5 5

缓存读取 2,432 8,576 8,576 9,088 8,576

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 20:03:25 ~ 20:03:40 (14.3s) 20:03:11 ~ 20:03:26 (14.1s) 20:03:09 ~ 20:03:29 (18.2s) 20:03:16 ~ 20:03:41 (24.4s) 20:03:14 ~ 20:03:29 (14.4s)

提示词

查看本步 Prompt

以下是对你自身的 metadata 探针。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。按以下 3 行回答自己的身份，每行一个 label，不要寒暄不要解释：
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) 1,704 ms 1,941 ms 2,848 ms 3,222 ms 975 ms

总耗时 3,448 ms 3,354 ms 7,039 ms 13,134 ms 3,758 ms

输入 188 1,693 157 466 157

输出 20 20 149 130 95

缓存读取 10,112 8,576 10,112 10,112 10,112

输出速度 - - 35.6 tok/s 13.1 tok/s 34.1 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 20:04:13 ~ 20:04:28 (14.0s) 20:03:52 ~ 20:04:52 (58.2s) 20:03:56 ~ 20:04:11 (11.7s) 20:04:17 ~ 20:04:42 (23.3s) 20:04:07 ~ 20:04:27 (15.2s)

提示词

查看本步 Prompt

请直接回答你的知识截止日期，格式 YYYY-MM，不要附加任何说明。

HTTP 头(TTFB) 1,697 ms 1,948 ms 997 ms 4,991 ms 1,411 ms

总耗时 2,657 ms 2,895 ms 999 ms 11,716 ms 4,480 ms

输入 232 193 8,136 1,737 193

输出 8 8 42 8 8

缓存读取 10,112 10,112 0 8,576 10,112

输出速度 - - burst - -

响应

2024-06

2024-06

接口已经正常响应。你可以继续用同一个脚本验证延迟、流式输出和客户端解析逻辑。

2024-06

2024-06

步骤 4: identity_free text

时间 20:05:01 ~ 20:05:21 (16.1s) 20:05:28 ~ 20:05:53 (21.4s) 20:04:39 ~ 20:04:54 (11.8s) 20:05:26 ~ 20:05:56 (26.8s) 20:05:05 ~ 20:05:30 (13.3s)

提示词

查看本步 Prompt

中文做个自我介绍：你是什么产品或助手？谁开发的？50 字内。

HTTP 头(TTFB) - 7,886 ms 906 ms 6,815 ms 949 ms

总耗时 - 10,479 ms 907 ms 13,854 ms 2,662 ms

输入 - 222 8,168 1,891 6,372

输出 - 27 42 22 36

缓存读取 - 10,112 0 8,576 3,968

输出速度 - - burst - 21.0 tok/s

响应

我是 Codex，OpenAI 开发的编程助手，可协助阅读、修改、调试代码并处理软件工程任务。

我是 OpenAI 的 AI 助手，当前作为 Codex 编程代理运行，由 OpenAI 开发。

接口已经正常响应。你可以继续用同一个脚本验证延迟、流式输出和客户端解析逻辑。

我是 Codex，基于 GPT-5 的编码助手，由 OpenAI 打造。

我是 OpenAI 开发的 Codex 编程助手，能在共享工作区中读代码、运行命令、协助修改与验证项目。

步骤 5: knowledge_recall text

时间 20:06:06 ~ 20:06:26 (16.5s) 20:06:33 ~ 20:06:53 (17.8s) 20:05:30 ~ 20:05:50 (16.8s) 20:06:21 ~ 20:06:46 (23.8s) 20:06:10 ~ 20:06:30 (15.5s)

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) 1,466 ms 3,089 ms 1,670 ms 3,885 ms 970 ms

总耗时 5,807 ms 6,961 ms 5,927 ms 12,656 ms 4,715 ms

输入 8,126 8,126 8,126 8,157 8,126

输出 179 179 179 182 182

缓存读取 0 0 0 0 0

输出速度 41.2 tok/s 46.2 tok/s 42.0 tok/s 20.8 tok/s 48.6 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":"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) 38,925 49,158 44,823 49,948 49,164

总输出 212 239 417 347 326

总缓存创建 0 0 0 0 0

总缓存读取 22,656 37,376 18,688 36,352 32,768

输出速度 41.2 tok/s 46.2 tok/s 38.8 tok/s 16.7 tok/s 38.0 tok/s

总执行时间 1m 18.7s 2m 6.8s 1m 11.7s 1m 57.1s 1m 10.8s

总墙钟时间

4m 1.5s

含步骤间等待

4m 25.0s

含步骤间等待

3m 23.9s

含步骤间等待

4m 24.0s

含步骤间等待

3m 60.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0990

OpenAI: GPT-5.5

$0.0848

OpenAI: GPT-5.5

$0.1525

OpenAI: GPT-5.5

$0.0966

OpenAI: GPT-5.5

$0.1081

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

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
cx-Zeus	via: 1.1 Caddy cf-ray: a14d85c4cf78dcee-LAX server: cloudflare x-oneapi-request-id: 20260702200607221368481LvOPSMuX	via_chain: ["1.1 Caddy"]	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-rayhdr:via	a14d85c4cf78dcee-LAX
O-Plus	cf-ray: a14d86715ea8fb9b-AMS server: cloudflare x-new-api-version: v1.0.0-rc.8 x-oneapi-request-id: 202607021206361284484498268d9d6F8q1c8G6	—	hdr:x-new-api-versionhdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-ray	a14d86715ea8fb9b-AMS
O-Pro	cf-ray: a14d84e6af362fda-AMS server: cloudflare	—	hdr:server:cloudflarehdr:cf-ray	a14d84e6af362fda-AMS
O-Api	x-new-api-version: 20260701-395ebb1 x-oneapi-request-id: 20260702120623235406363TyhGseRQ	—	hdr:x-new-api-versionhdr:x-oneapi-request-id	—
官方基准	server: nginx/1.22.1	—	—	—

协议指纹评分（机器自动）

总分 / 100 99 91 85 74 100

基线状态已对照基线已对照基线已对照基线已对照基线已对照基线

缓存命中比 10.0 /10 ×20 10.0 /10 ×20 6.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

延迟基线 9.2 /10 ×15 4.8 /10 ×15 9.8 /10 ×15 0.0 /10 ×15 10.0 /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 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8 10.0 /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 0.0 /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 99 v3.26.0 91 v3.26.0 85 v3.26.0 74 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 3785 ms 5654 ms 3433 ms 11817 ms 3471 ms

输出速度 41.2 tok/s 46.2 tok/s 38.8 tok/s 16.7 tok/s 38.0 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 ⓘ

$0.0990

$0.0848

$0.1525

$0.0966

$0.1081

缓存读取占比 ⓘ 58% 76% 42% 73% 67%

揭晓

通道

服务商 PackyCode ↗

通道cx-Zeus

服务商 LyraAPI ↗

通道O-Plus

服务商 ModelFlare ↗

通道O-Pro

服务商 CCAPI ↗

通道O-Api

官方基准

返回列表