模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

前 5 步同一会话，第 6 步独立新会话，覆盖 6 个独立信号： 1. ping/pong —— 单字指令遵循 + 建立缓存上下文。 2. 身份（结构化）—— 三行 vendor / brand / model 格式，机器可解析。 3. 知识截止 —— 跨越 5 分钟 cache 边界后追问，检测 sliding 5m cache 是否真正命中。 4. 身份（自由格式）—— 自然语言自报身份，给包装层（Kiro 等）暴露品牌的机会。 5. 世界知识层级 —— 5 道公共事件硬事实题，按答对档位映射实测 tier，与请求模型对照。 6. 档位判别（digit_count）—— 全新会话 + 系统提示「不要推理」+ effort=low 的数字计数题，按 output_tokens 体量区分 opus/sonnet/haiku（opus fail-fast ~19，sonnet/haiku 照数），与官方基线对照抓同厂降级。前 5 步之间随机延迟 1–4 分钟。第 3 步（cutoff）累计跨度 > 5 分钟保住 sliding cache 检测；总跨度 > 6 分钟。第 6 步是独立新会话（不复用前 5 轮上下文，否则 opus 不再 fail-fast）。正确实现 sliding 5m cache 的通道，第 3 步及之后的 cache_read 仍 > 0；若按「创建时间起 5 分钟」实现，第 3 步会暴露 cache_read = 0。

5 通道 · 6 步骤

2026/06/28 运行日期

claude-sonnet-4-6 官方基准

2 任务失败执行不完整

已揭晓揭晓进度

步骤 / 指标

SSSAiCode

O-Max-L2

WorldBase.ai

M-

官方基准

SSSAiCode

O-Max

SudoCode

M-Max

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6

客户端 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160 claude-cli v2.1.160

偏差标红基于官方基准

步骤 1: ping text

时间 10:39:36 ~ 10:39:41 (4.2s) 10:39:42 ~ 10:39:52 (6.8s) 10:39:51 ~ 10:40:06 (4.9s) 10:39:34 ~ 10:39:39 (3.0s) 10:39:46 ~ 10:39:51 (4.5s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,932 ms 5,112 ms 3,718 ms 1,858 ms 3,126 ms

首文本(SSE delta) 3,034 ms 5,113 ms 3,719 ms 1,956 ms 3,129 ms

总耗时 3,150 ms 5,226 ms 3,877 ms 2,034 ms 3,307 ms

输入 3 1 3 3 3

输出 5 2 5 5 5

缓存创建

31,987

5m: 0 / 1h: 31,987

21,970

5m: 0 / 1h: 21,970

32,318

5m: 0 / 1h: 32,318

31,985

5m: 0 / 1h: 31,985

32,318

5m: 0 / 1h: 32,318

缓存读取 0 14,122 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 10:43:16 ~ 10:43:21 (3.7s) 10:43:32 ~ 10:43:42 (8.0s) 10:41:15 ~ 10:41:20 (3.8s) 10:41:38 ~ 10:41:43 (3.5s) 10:40:51 ~ 10:40:56 (4.3s)

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。自报身份，输出严格遵守以下 3 行格式（不允许 markdown、代码块、额外文本）：
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) 2,470 ms 4,878 ms 2,327 ms 2,091 ms 2,945 ms

首文本(SSE delta) 2,715 ms 6,663 ms 2,328 ms 2,362 ms 2,946 ms

总耗时 2,730 ms 6,663 ms 2,694 ms 2,503 ms 3,347 ms

输入 3 67 3 3 3

输出 22 35 22 22 22

缓存创建

192

5m: 0 / 1h: 192

5,623

5m: 0 / 1h: 5,623

192

5m: 0 / 1h: 192

192

5m: 0 / 1h: 192

192

5m: 0 / 1h: 192

缓存读取 31,987 38,920 32,318 31,985 32,318

输出速度 - burst - - -

响应

vendor: anthropic
brand: claude
model: claude-sonnet-4-5

I'm Claude Code, powered by Anthropic's Claude model — specifically claude-sonnet-4-6.

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-5

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

步骤 3: cutoff text

时间 10:46:14 ~ 10:46:19 (2.6s) 10:46:06 ~ 10:46:16 (5.7s) 10:44:56 ~ 10:45:01 (3.1s) 10:43:14 ~ 10:43:19 (3.3s) 10:43:00 ~ 10:43:05 (4.4s)

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) 1,513 ms 4,311 ms 1,928 ms 1,624 ms 3,104 ms

首文本(SSE delta) 1,734 ms 4,312 ms 2,036 ms 1,808 ms 3,105 ms

总耗时 1,802 ms 4,582 ms 2,115 ms 1,878 ms 3,354 ms

输入 3 67 3 3 3

输出 8 5 8 8 8

缓存创建

5m: 0 / 1h: 57

5,659

5m: 0 / 1h: 5,659

5m: 0 / 1h: 57

5m: 0 / 1h: 55

缓存读取 32,179 38,943 32,510 32,177 32,510

响应

2025-04

2025-08

2025-08

2025-04

2025-08

步骤 4: identity_free text

时间 10:49:47 ~ 10:49:52 (3.3s) 10:49:29 ~ 10:49:44 (11.4s) 10:48:30 ~ 10:48:40 (6.8s) 10:46:45 ~ 10:46:50 (4.0s) 10:46:48 ~ 10:46:58 (5.1s)

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) 1,974 ms 8,907 ms 5,552 ms 1,877 ms 3,308 ms

首文本(SSE delta) 2,411 ms 9,277 ms 5,554 ms 2,986 ms 3,386 ms

总耗时 2,471 ms 9,968 ms 5,822 ms 3,125 ms 4,078 ms

输入 3 67 3 3 3

输出 47 74 26 62 50

缓存创建

5m: 0 / 1h: 35

5,686

5m: 0 / 1h: 5,686

5m: 0 / 1h: 35

5m: 0 / 1h: 43

缓存读取 32,236 38,961 32,567 32,232 32,565

输出速度 - 107.1 tok/s - - 72.3 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。Anthropic 是一家专注于 AI 安全研究的美国公司。

我是 Claude Code，这是 Anthropic 开发的 Claude 模型在命令行中的交互工具。我能帮你完成软件工程任务、编写代码、调试程序等工作。

我是 Claude，由 Anthropic 公司开发的 AI 助手。

我是 Claude，由 Anthropic 开发的 AI 助手。我被设计用于对话、分析、编程、写作等多种任务，致力于提供安全、有益的帮助。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助回答问题、分析代码、写作、推理等多种任务。

步骤 5: knowledge_recall text

时间 10:53:31 ~ 10:53:46 (12.3s) 10:51:47 ~ 10:51:57 (6.4s) 10:50:12 ~ 10:50:17 (4.5s) 10:49:21 ~ 10:49:26 (4.8s) 10:50:53 ~ 10:51:03 (5.3s)

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) 8,037 ms 3,237 ms 1,552 ms 1,654 ms 1,918 ms

首文本(SSE delta) 11,433 ms 3,238 ms 1,553 ms 3,878 ms 1,920 ms

总耗时 11,595 ms 5,258 ms 3,490 ms 4,015 ms 4,338 ms

输入 3 1 3 3 3

输出 198 282 204 198 205

缓存创建

7,707

5m: 0 / 1h: 7,707

7,595

5m: 0 / 1h: 7,595

8,017

5m: 0 / 1h: 8,017

7,726

5m: 0 / 1h: 7,726

8,047

5m: 0 / 1h: 8,047

缓存读取 0 4,882 0 0 0

输出速度 - 139.6 tok/s 105.3 tok/s - 84.8 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

```json
{"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

{"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

步骤 6: digit_count text

时间 10:55:14 ~ 10:55:19 (1.1s) 10:54:13 ~ 10:54:33 (15.6s) 10:54:02 ~ 10:54:27 (21.4s) 10:51:27 ~ 10:51:32 (0.9s) 10:53:51 ~ 10:54:11 (19.1s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=148048600788961458258385774443351775956714474291028826940170427955128022

HTTP 头(TTFB) - 2,081 ms 2,931 ms - 1,836 ms

首文本(SSE delta) - 2,097 ms 2,932 ms - 1,857 ms

总耗时 - 14,541 ms 20,116 ms - 18,142 ms

输入 - 473 164 - 164

输出 - 731 1,808 - 1,246

缓存创建

5m: 0 / 1h: 0

缓存读取 - 0 0 - 0

输出速度 - 58.7 tok/s 105.2 tok/s - 76.5 tok/s

响应

失败

评测失败

```json
{"n":72,"c":[7,9,5]}
```

## Counting digits in:
`148048600788961458258385774443351775956714474291028826940170427955128022`

### Step 1 — Length (n)

Splitting into groups of 10:
```
1480486007  → 10
8896145825  → 10
838577444

失败

评测失败

## Counting digits in s

**s** = `148048600788961458258385774443351775956714474291028826940170427955128022`

### Length
Splitting into groups of 10:
```
1480486007 | 8896145825 | 8385774443 | 35177595

总计

总输入(含缓存)

136,395

仅前 6 步

183,037

138,193

136,410

仅前 6 步

138,227

总输出

280

仅前 6 步

1,129

2,073

295

仅前 6 步

1,536

总缓存创建

39,978

5m: 0 / 1h: 39,978

仅前 6 步

46,533

5m: 0 / 1h: 46,533

40,619

5m: 0 / 1h: 40,619

40,001

5m: 0 / 1h: 40,001

仅前 6 步

40,655

5m: 0 / 1h: 40,655

总缓存读取

96,402

仅前 6 步

135,828

97,395

96,394

仅前 6 步

97,393

输出速度

仅前 6 步

71.7 tok/s

105.2 tok/s

仅前 6 步

77.4 tok/s

总执行时间

27.2s

仅前 6 步

53.8s

44.4s

19.5s

仅前 6 步

42.6s

总墙钟时间

15m 43.2s

含步骤间等待

仅前 6 步

14m 51.0s

含步骤间等待

14m 35.6s

含步骤间等待

11m 58.0s

含步骤间等待

仅前 6 步

14m 24.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1831

Anthropic: Claude Sonnet 4.6 (部分)

$0.2342

Anthropic: Claude Sonnet 4.6

$0.2132

Anthropic: Claude Sonnet 4.6

$0.1834

Anthropic: Claude Sonnet 4.6 (部分)

$0.2053

Anthropic: Claude Sonnet 4.6

通道指纹

来源步骤 knowledge_recall digit_count digit_count knowledge_recall digit_count

平台 direct (100%) direct (100%) direct (100%) direct (100%) one-api (100%)

上游 Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN - cloudflare - - -

ID 格式 msg_019L9ATbNfycu1cjjYrSBMXv msg_f9545291b5ad4188aceeae629ec1c6f2 msg_016M6a1tg7a8me45sYDbpyss msg_01MMu3Pog7ZUgxdsfaYe1fBs msg_01JKhAM8J13vfG8K2S9sSc5f

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-Max-L2	cf-ray: a12966d98b962edf-LAX server: Photon-Edge set-cookie: _cfuvid=l8e68zwVjqP7k6AYpLTViL1GVxTDG6ihDbF05D7pdMc-1782615213.0506053-1.0.1.1-34z.UDdbR7o2.VUagQroP.9SXSokACu4cUIbQYZ4GQI; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.anthropic.com	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1782615212939-32ay59w0tn0sz req_011CcUuGjUfofkQ6Nr23BkPV a12966d98b962edf-LAX
M-	cf-ray: a12967dfbda6210f-AMS server: cloudflare	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	74731202-1d25-44dd-bbf0-6a1e6627e76a a12967dfbda6210f-AMS
官方基准	server: nginx/1.22.1	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcUuK8YYgLRMVVbT9xiSY
O-Max	cf-ray: a12960be0f3ed938-LAX server: Photon-Edge set-cookie: _cfuvid=Yf7ocQ_gKQYjErIWelkS.gLkz5wB2uiYt2f5J9f9QAw-1782614962.8878603-1.0.1.1-ylCTYGYdpQotN0WbttLPbdEmU98JXkMrzdI6jBO7tFc; HttpOnly; SameSite=None; Secure; Path=/; Domain=api.anthropic.com	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1782614962776-hlnutbd11ifao req_011CcUtxHrYmUbsoH3yLGnxk a12960be0f3ed938-LAX
M-Max	server: nginx x-new-api-version: 0b8be5cf x-oneapi-request-id: 202606281053524725985214oyyxU4w	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—

协议指纹评分（机器自动）

总分 / 100 86 100 98

基线状态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

缓存连续性 — 10.0 /10 ×14 10.0 /10 ×14 — 10.0 /10 ×14

sliding 5m — 10.0 /10 ×13 10.0 /10 ×13 — 10.0 /10 ×13

缓存 TTL 一致 — 10.0 /10 ×15 10.0 /10 ×15 — 10.0 /10 ×15

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

身份(结构) — 0.0 /10 ×7 10.0 /10 ×7 — 10.0 /10 ×7

知识截止 — 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

原生 msg-ID — 10.0 /10 ×8 10.0 /10 ×8 — 10.0 /10 ×8

Req-ID 透传 — 0.0 /10 ×4 10.0 /10 ×4 — 0.0 /10 ×4

stop_reason — 10.0 /10 ×3 10.0 /10 ×3 — 10.0 /10 ×3

service_tier — 10.0 /10 ×6 10.0 /10 ×6 — 10.0 /10 ×6

inference_geo — 10.0 /10 ×5 10.0 /10 ×5 — 10.0 /10 ×5

SDK 一致 — 10.0 /10 ×2 10.0 /10 ×2 — 10.0 /10 ×2

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

延迟基线 — 7.4 /10 ×5 10.0 /10 ×5 — 10.0 /10 ×5

流式投递 — 10.0 /10 ×5 10.0 /10 ×5 — 10.0 /10 ×5

综合结论

与基线相似度执行失败，不评分 86 v3.25.0 基准（参考）执行失败，不评分 98 v3.25.0

雷达图例

共 21 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 SDK 一致 sdk_consistency
4 流式投递 buffer_dump_match
5 自由身份 identity_free_clean
6 停止原因 stop_reason_present
7 系统提示纯净 system_prompt_clean
8 Service Tier service_tier_present
9 缓存命中比 cache_hit_ratio_match
10 缓存 TTL 一致 cache_ttl_consistency
11 推理地区 inference_geo_present
12 缓存连续 cache_continuity_intra
13 知识召回 knowledge_recall_match
14 延迟基线 latency_baseline_match
15 原生 msg-ID anthropic_msg_id_format
16 Sliding 缓存 cache_sliding_correctness
17 结构化身份 identity_structured_match
18 自报身份一致 self_identity_consistency
19 信封自报一致 envelope_self_report_match
20 档位思考量 tier_thinking_volume_match
21 Req-ID 透传 anthropic_request_id_passthrough

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 4350 ms 7706 ms 6352 ms 2711 ms 6094 ms

输出速度 - 71.7 tok/s 105.2 tok/s - 77.4 tok/s

完成步骤 5/6 成功 · 1 失败 6/6 成功 6/6 成功 5/6 成功 · 1 失败 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1831 (部分)

$0.2342

$0.2132

$0.1834 (部分)

$0.2053

缓存读取占比 ⓘ 71% (部分) 74% 70% 71% (部分) 70%

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 WorldBase.ai ↗

通道M-

官方基准

服务商 SSSAiCode ↗

通道O-Max

服务商 SudoCode ↗

通道M-Max

返回列表