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

3 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

JikeAI

AnyAIGC

M-Pro

right

u-cx

官方基准

right

o-cx

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 22:01:23 ~ 22:01:38 (14.9s) 22:01:21 ~ 22:01:36 (10.0s) 22:01:27 ~ 22:01:32 (3.0s) 22:01:29 ~ 22:01:44 (12.5s) 22:01:25 ~ 22:01:40 (12.9s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,665 ms 7,734 ms 573 ms 1,052 ms 573 ms

总耗时 4,078 ms 8,825 ms 2,463 ms 1,856 ms 2,488 ms

输入 6,338 1,699 1,020 6,819 1,276

输出 5 5 869 5 808

缓存读取 2,432 7,040 5,376 1,920 5,120

输出速度 - - burst - burst

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 22:02:01 ~ 22:02:21 (18.8s) 22:02:17 ~ 22:02:42 (20.5s) 22:02:06 ~ 22:02:26 (15.7s) 22:02:29 ~ 22:02:39 (8.3s) 22:02:12 ~ 22:02:32 (15.2s)

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) 6,215 ms 7,947 ms 757 ms 900 ms 675 ms

总耗时 7,199 ms 9,465 ms 5,217 ms 7,550 ms 4,639 ms

输入 327 6,952 906 296 1,290

输出 306 185 1,342 344 1,231

缓存读取 8,576 1,920 5,632 8,576 5,248

输出速度 burst 121.9 tok/s burst 51.7 tok/s burst

响应

vendor: openai
brand: gpt-5
model: gpt-5

vendor: openai
brand: codex
model: gpt-5

vendor: openai
brand: gpt
model: gpt-5

vendor: openai
brand: gpt
model: gpt-5

vendor: openai
brand: gpt
model: gpt-5

步骤 3: cutoff text

时间 22:02:56 ~ 22:03:21 (20.6s) 22:03:04 ~ 22:03:24 (19.1s) 22:02:51 ~ 22:03:06 (12.8s) 22:03:06 ~ 22:03:21 (12.2s) 22:03:16 ~ 22:03:31 (12.6s)

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) 6,135 ms 6,949 ms 821 ms 879 ms 994 ms

总耗时 9,554 ms 7,766 ms 2,286 ms 1,496 ms 2,194 ms

输入 375 340 1,342 340 1,574

输出 399 27 821 8 32

缓存读取 8,576 8,576 5,248 8,576 7,552

输出速度 116.7 tok/s - burst - 26.7 tok/s

响应

2024-06

2024-08

2024-06

2024-06

2024-06

步骤 4: identity_free text

时间 22:03:40 ~ 22:04:00 (19.2s) 22:04:00 ~ 22:04:20 (16.3s) 22:03:29 ~ 22:03:49 (16.3s) 22:03:57 ~ 22:04:12 (13.0s) 22:04:11 ~ 22:04:26 (13.1s)

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) 4,541 ms 4,415 ms 625 ms 978 ms 710 ms

总耗时 5,700 ms 5,211 ms 5,785 ms 2,331 ms 2,579 ms

输入 404 372 614 371 1,249

输出 141 16 817 31 813

缓存读取 8,576 8,576 6,016 8,576 5,376

输出速度 121.7 tok/s - 158.3 tok/s 22.9 tok/s burst

响应

我是 OpenAI 的 GPT-5，由 OpenAI 开发。

我是 Codex 编码助手，由 OpenAI 开发。

我是 OpenAI 开发的 GPT-5 助手，可用于问答、写作与编程。

我叫 GPT-5，属于 OpenAI API，由 OpenAI 开发。

我是 OpenAI 开发的 GPT 助手，可协助写作、编程和问答。

步骤 5: knowledge_recall text

时间 22:04:39 ~ 22:04:59 (18.4s) 22:04:57 ~ 22:05:17 (19.4s) 22:04:30 ~ 22:04:50 (18.7s) 22:04:53 ~ 22:05:13 (15.6s) 22:04:51 ~ 22:05:11 (19.4s)

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) 5,125 ms 6,704 ms 955 ms 872 ms 1,083 ms

总耗时 6,710 ms 8,381 ms 8,280 ms 4,862 ms 8,958 ms

输入 6,772 6,741 7,037 6,741 7,037

输出 185 179 383 179 404

缓存读取 0 0 0 0 0

输出速度 116.7 tok/s 106.7 tok/s 52.3 tok/s 44.9 tok/s 51.3 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) 42,376 42,216 33,191 42,215 35,722

总输出 1,036 412 4,232 567 3,288

总缓存创建 0 0 0 0 0

总缓存读取 28,160 26,112 22,272 27,648 23,296

输出速度 117.6 tok/s 113.9 tok/s 96.1 tok/s 46.2 tok/s 48.0 tok/s

总执行时间 1m 31.9s 1m 25.2s 1m 6.6s 1m 1.6s 1m 13.2s

总墙钟时间

3m 35.9s

含步骤间等待

3m 55.9s

含步骤间等待

3m 22.9s

含步骤间等待

3m 43.9s

含步骤间等待

3m 45.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0581

OpenAI: GPT-5.4

$0.0530

OpenAI: GPT-5.4

$0.0963

OpenAI: GPT-5.4

$0.0518

OpenAI: GPT-5.4

$0.0862

OpenAI: GPT-5.4

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%) 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	cf-ray: a14e33773adc71b2-FRA server: cloudflare set-cookie: p_uv_id=15ccd41c092be1576574a254ef9dcacd; path=/; expires=Fri, 03-Jul-26 00:00:00 GMT; HttpOnly x-oneapi-request-id: 202607021404419310776098268d9d6Reo1MpFI	—	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-ray	a14e33773adc71b2-FRA
M-Pro	server: nginx	—	—	—
u-cx	via: 1.1 Caddy cf-ray: a14e33360906fd4f-NRT server: cloudflare	via_chain: ["1.1 Caddy"]	hdr:server:cloudflarehdr:cf-rayhdr:via	a14e33360906fd4f-NRT
官方基准	server: nginx/1.22.1	—	—	—
o-cx	via: 1.1 Caddy cf-ray: a14e33b9dcebfd4f-NRT server: cloudflare	via_chain: ["1.1 Caddy"]	hdr:server:cloudflarehdr:cf-rayhdr:via	a14e33b9dcebfd4f-NRT

协议指纹评分（机器自动）

总分 / 100 84 85 84 100 84

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 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

延迟基线 2.3 /10 ×15 1.3 /10 ×15 9.9 /10 ×15 10.0 /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 0.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8

身份(结构) 6.7 /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 84 v3.26.0 85 v3.26.0 84 v3.26.0 基准（参考） 84 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 6648 ms 7930 ms 4806 ms 3619 ms 4172 ms

输出速度 117.6 tok/s 113.9 tok/s 96.1 tok/s 46.2 tok/s 48.0 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 ⓘ

$0.0581

$0.0530

$0.0963

$0.0518

$0.0862

缓存读取占比 ⓘ 66% 62% 67% 65% 65%

揭晓

通道

服务商 JikeAI ↗

通道cx

服务商 AnyAIGC ↗

通道M-Pro

服务商 right ↗

通道u-cx

官方基准

服务商 right ↗

通道o-cx

返回列表