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

2026/06/22 运行日期

claude-sonnet-4-6 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

LingxiCode

M-

Aiide

O-Max

luminai

M-Kiro/Antigravity

SSSAiCode

R-Kiro

官方基准

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 08:48:03 ~ 08:48:18 (13.4s) 08:48:01 ~ 08:48:26 (22.2s) 08:47:59 ~ 08:48:09 (6.3s) 08:48:05 ~ 08:48:10 (3.9s) 08:48:07 ~ 08:48:37 (25.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 11,346 ms 20,454 ms 4,899 ms 2,818 ms 24,319 ms

首文本(SSE delta) 11,348 ms 20,493 ms 4,962 ms 2,903 ms 24,321 ms

总耗时 11,531 ms 20,580 ms 5,035 ms 2,968 ms 24,510 ms

输入 3 3 1,443 0 3

输出 5 5 1 2 5

缓存创建

32,318

5m: 0 / 1h: 32,318

32,457

5m: 0 / 1h: 32,457

4,111

5m: 4,111 / 1h: 0

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 23,299 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 08:49:54 ~ 08:49:59 (4.3s) 08:51:32 ~ 08:51:57 (22.1s) 08:52:02 ~ 08:52:07 (4.7s) 08:52:07 ~ 08:52:22 (11.3s) 08:51:07 ~ 08:51:22 (12.5s)

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) 3,185 ms 19,599 ms 3,444 ms 7,013 ms 11,048 ms

首文本(SSE delta) 3,186 ms 19,600 ms 3,450 ms 7,015 ms 11,050 ms

总耗时 3,515 ms 19,926 ms 3,701 ms 7,296 ms 11,483 ms

输入 3 3 1,453 0 3

输出 22 22 24 17 22

缓存创建

202

5m: 0 / 1h: 202

1,666

5m: 0 / 1h: 1,666

4,139

5m: 4,139 / 1h: 0

202

5m: 0 / 1h: 202

缓存读取 32,318 30,993 23,455 0 32,318

响应

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

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

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

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

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

步骤 3: cutoff text

时间 08:53:28 ~ 08:53:33 (4.1s) 08:55:33 ~ 08:55:43 (5.8s) 08:53:08 ~ 08:53:48 (35.3s) 08:54:31 ~ 08:54:36 (3.5s) 08:54:37 ~ 08:54:47 (7.0s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,954 ms 3,953 ms 33,919 ms 2,466 ms 4,976 ms

首文本(SSE delta) 2,955 ms 3,955 ms 34,120 ms 2,606 ms 5,052 ms

总耗时 3,163 ms 4,525 ms 34,178 ms 2,613 ms 6,009 ms

输入 3 3 1,456 0 3

输出 8 8 4 3 8

缓存创建

5m: 0 / 1h: 38

32,716

5m: 0 / 1h: 32,716

4,148

5m: 4,148 / 1h: 0

5m: 0 / 1h: 55

缓存读取 32,520 0 23,506 0 32,520

响应

2025-08

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 08:55:22 ~ 08:55:27 (4.8s) 08:59:34 ~ 08:59:44 (8.9s) 08:55:44 ~ 08:55:54 (6.1s) 08:57:12 ~ 08:57:17 (4.5s) 08:56:16 ~ 08:56:31 (12.2s)

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) 3,186 ms 6,856 ms 4,208 ms 2,299 ms 10,337 ms

首文本(SSE delta) 3,187 ms 6,857 ms 4,767 ms 2,473 ms 10,338 ms

总耗时 3,807 ms 7,740 ms 4,866 ms 3,686 ms 11,164 ms

输入 3 3 1,458 0 3

输出 49 46 50 41 50

缓存创建

5m: 0 / 1h: 43

5m: 0 / 1h: 41

4,154

5m: 4,154 / 1h: 0

5m: 0 / 1h: 41

缓存读取 32,558 32,716 23,544 0 32,575

输出速度 79.0 tok/s 52.1 tok/s - 33.8 tok/s 60.5 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助您完成写作、分析、编程、问答等多种任务。

我叫 Claude，是由 Anthropic 开发的 AI 助手，当前使用的模型版本为 claude-sonnet-4-6。

我是 Claude Sonnet 4.6，由 Anthropic 开发的 AI 助手，专注于帮助用户完成编程、写作、分析等各类任务。

我是 Kiro，一款由 Anthropic 提供 AI 能力的智能开发环境助手，基于 Claude 模型构建，专为软件开发任务设计，帮助开发者更高效地编写、调试和理解代码。

我叫 Claude，是 Anthropic 开发的 AI 助手，属于 Claude 产品系列。当前版本为 Claude Sonnet 4.6。

步骤 5: knowledge_recall text

时间 08:57:14 ~ 08:57:24 (6.7s) 09:01:13 ~ 09:03:33 (136.9s) 08:59:41 ~ 08:59:51 (5.8s) 08:58:38 ~ 08:58:43 (4.6s) 08:58:15 ~ 08:58:25 (8.4s)

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) 3,377 ms 21,596 ms 2,576 ms 1,785 ms 3,981 ms

首文本(SSE delta) 3,378 ms 21,597 ms 4,452 ms 1,855 ms 4,005 ms

总耗时 5,803 ms 23,986 ms 4,525 ms 3,702 ms 7,395 ms

输入 3 3 376 0 3

输出 205 205 210 221 205

缓存创建

8,039

5m: 0 / 1h: 8,039

8,053

5m: 0 / 1h: 8,053

1,069

5m: 1,069 / 1h: 0

8,055

5m: 0 / 1h: 8,055

缓存读取 0 0 6,063 0 0

输出速度 84.5 tok/s 85.8 tok/s burst 119.7 tok/s 60.5 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

步骤 6: digit_count text

时间 08:59:11 ~ 08:59:26 (11.6s) 09:07:02 ~ 09:07:52 (46.7s) 09:01:27 ~ 09:01:32 (3.7s) 09:02:32 ~ 09:02:47 (12.3s) 09:01:25 ~ 09:01:55 (27.6s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=895440571125438363295428925836099461174583944440134293033394882565058199

HTTP 头(TTFB) 2,024 ms 9,409 ms 2,106 ms 1,732 ms 8,054 ms

首文本(SSE delta) 2,025 ms 9,800 ms 2,334 ms 1,761 ms 8,056 ms

总耗时 10,804 ms 27,148 ms 2,403 ms 11,297 ms 26,493 ms

输入 164 164 189 0 164

输出 757 1,511 14 796 1,439

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 - 0 0

输出速度 86.2 tok/s 87.1 tok/s - 83.5 tok/s 78.0 tok/s

响应

```json
{"n":72,"c":[5,2,10]}
```

**Verification:**

| Digit | Positions (1-indexed) | Count |
|---|---|---|
| **0** | 6, 31, 48, 55, 67 | **5** |
| **7** | 8, 38 | **2** |
| **9** | 2, 20, 25, 32, 3

```python
s = "895440571125438363295428925836099461174583944440134293033394882565058199"
n = len(s)
c = [s.count('0'), s.count('7'), s.count('9')]
# n=72, c=[5,2,10]
```

**Digit-by-digit verification

{"n":71,"c":[3,1,7]}

```json
{"n":72,"c":[5,2,10]}
```

## Counting digits in s

**s** = `895440571125438363295428925836099461174583944440134293033394882565058199`

### Length Count (groups of 10)
```
8954405711  → 10
2543836329  → 10
5428925836  → 10
0994

总计

总输入(含缓存) 138,215 138,821 123,863 0 138,263

总输出 1,046 1,797 303 1,080 1,729

总缓存创建

40,640

5m: 0 / 1h: 40,640

74,933

5m: 0 / 1h: 74,933

17,621

5m: 17,621 / 1h: 0

40,671

5m: 0 / 1h: 40,671

总缓存读取 97,396 63,709 99,867 0 97,413

输出速度 85.5 tok/s 85.5 tok/s burst 84.0 tok/s 74.8 tok/s

总执行时间 45.0s 4m 2.5s 1m 2.0s 39.9s 1m 33.3s

总墙钟时间

11m 23.0s

含步骤间等待

19m 50.4s

含步骤间等待

13m 33.0s

含步骤间等待

14m 42.0s

含步骤间等待

13m 48.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1978

Anthropic: Claude Sonnet 4.6

$0.3276

Anthropic: Claude Sonnet 4.6

$0.0949

Anthropic: Claude Sonnet 4.6

>=$0.0162

Anthropic: Claude Sonnet 4.6 (成本不完整)

$0.2082

Anthropic: Claude Sonnet 4.6

通道指纹

来源步骤 digit_count digit_count digit_count digit_count digit_count

平台 one-api (100%) one-api (100%) one-api (100%) direct (75%) direct (100%)

上游 Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (75%) Anthropic 直连 (100%)

CDN cloudflare cloudflare cloudflare - -

ID 格式 msg_015TKGBmFrMDDPuQi9BXykti msg_01AdFTDnYgg2WAAwzu9r5si5 msg_01VBpS2SZ5t4TQqqJbRGg38U57 msg_03408fc52e7f4e919b4006659a5695ff msg_01HdtKYzBJabE5uT66wx6VUQ

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-	cf-ray: a0f74f1bbd6fe393-NRT server: cloudflare set-cookie: server_name_session=3de609b8cf2a7e28d141c3b6c8b41f57; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260622005913225906270c24okJS1	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-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f74f1bbd6fe393-NRT
O-Max	cf-ray: a0f75b0cae842d31-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 202606220107217831704268268d9d6PEsEXEVd	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-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f75b0cae842d31-AMS
M-Kiro/Antigravity	cf-ray: a0f75271cddcf5e2-AMS server: cloudflare x-oneapi-request-id: 202606220101293457369738268d9d6h3e6mIXA	id_format: msg_prefix extra_fields: ["container"] has_service_tier: true has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f75271cddcf5e2-AMS
R-Kiro	server: Photon-Edge x-amzn-requestid: 59949ece-6d83-40ce-a9cb-bdc0f7dbabc1	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.cache_fieldsupstream:x-amzn-requestidupstream:anthropic-direct-from-body	1782090153914-vwq619hmaenlg 59949ece-6d83-40ce-a9cb-bdc0f7dbabc1
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_011CcHPsX2ztEAdU5rdt71iE

协议指纹评分（机器自动）

总分 / 100 98 88 74 52 95

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 5.0 /10 ×20 10.0 /10 ×20 0.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 0.0 /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

缓存连续性 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 5.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 7.7 /10 ×13 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13

原生 msg-ID 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8

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

知识截止 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 0.0 /10 ×7 10.0 /10 ×7

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 0.0 /10 ×6 10.0 /10 ×6

inference_geo 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

延迟基线 10.0 /10 ×5 0.7 /10 ×5 9.3 /10 ×5 10.0 /10 ×5 0.0 /10 ×5

流式投递 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

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

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

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

综合结论

与基线相似度 98 v3.24.0 88 v3.24.0 74 v3.24.0 52 v3.24.0 基准（参考）

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 信封自报一致 envelope_self_report_match
19 中文遵循 instruction_following_lang
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 6437 ms 17318 ms 9118 ms 5260 ms 14509 ms

输出速度 85.5 tok/s 85.5 tok/s burst 84.0 tok/s 74.8 tok/s

完成步骤 6/6 成功 6/6 成功 6/6 成功 6/6 成功 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1978

$0.3276

$0.0949

>=$0.0162 (成本不完整)

$0.2082

缓存读取占比 ⓘ 70% 46% 81% - 70%

揭晓

通道

服务商 LingxiCode ↗

通道M-

服务商 Aiide ↗

通道O-Max

服务商 luminai ↗

通道M-Kiro/Antigravity

服务商 SSSAiCode ↗

通道R-Kiro

官方基准

返回列表