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

1 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

AIMZ

M-Max

LingxiCode

M-

SSSAiCode

R-Kiro

SudoCode

M-Max

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 12:36:57 ~ 12:37:07 (5.3s) 12:36:55 ~ 12:37:05 (5.1s) 12:37:01 ~ 12:37:06 (3.5s) 12:36:59 ~ 12:37:04 (4.1s) 12:37:03 ~ 12:37:13 (5.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,939 ms 3,544 ms 2,295 ms 2,722 ms 4,279 ms

首文本(SSE delta) 3,960 ms 3,548 ms 2,373 ms 2,724 ms 4,280 ms

总耗时 4,059 ms 3,954 ms 2,374 ms 2,894 ms 4,593 ms

输入 3 3 0 3 3

输出 5 5 2 5 5

缓存创建

31,985

5m: 0 / 1h: 31,985

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 12:38:37 ~ 12:38:42 (3.8s) 12:40:57 ~ 12:41:02 (4.7s) 12:40:26 ~ 12:40:36 (5.2s) 12:40:08 ~ 12:40:18 (5.3s) 12:40:10 ~ 12:40:20 (6.8s)

提示词

查看本步 Prompt

输出 3 行 metadata 描述你自己（当前 API 模型）。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。禁止任何额外字符：
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) 2,453 ms 3,067 ms 2,347 ms 3,974 ms 5,331 ms

首文本(SSE delta) 2,715 ms 3,068 ms 2,491 ms 3,976 ms 5,333 ms

总耗时 2,765 ms 3,640 ms 2,638 ms 4,376 ms 5,715 ms

输入 3 3 0 3 3

输出 22 22 18 22 22

缓存创建

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

缓存读取 31,985 32,318 0 32,318 32,318

响应

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

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 12:40:12 ~ 12:40:17 (3.8s) 12:42:30 ~ 12:42:35 (4.9s) 12:41:47 ~ 12:41:52 (3.3s) 12:43:19 ~ 12:43:24 (4.0s) 12:44:00 ~ 12:44:05 (4.7s)

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) 2,469 ms 3,670 ms 2,295 ms 2,731 ms 3,510 ms

首文本(SSE delta) 2,595 ms 3,671 ms 2,472 ms 2,732 ms 3,658 ms

总耗时 2,670 ms 4,033 ms 2,626 ms 2,972 ms 3,690 ms

输入 3 3 0 3 3

输出 8 8 3 8 8

缓存创建

5m: 0 / 1h: 57

5m: 0 / 1h: 38

5m: 0 / 1h: 55

5m: 0 / 1h: 57

缓存读取 32,174 32,507 0 32,507 32,507

响应

2025-04

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 12:41:33 ~ 12:41:38 (4.3s) 12:44:11 ~ 12:44:21 (5.6s) 12:45:00 ~ 12:45:05 (4.2s) 12:47:30 ~ 12:47:35 (4.8s) 12:47:59 ~ 12:48:14 (11.2s)

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) 2,925 ms 3,083 ms 2,223 ms 2,969 ms 9,414 ms

首文本(SSE delta) 3,085 ms 3,084 ms 2,531 ms 2,971 ms 9,812 ms

总耗时 3,085 ms 4,584 ms 3,436 ms 3,842 ms 10,196 ms

输入 3 3 0 3 3

输出 24 62 39 51 44

缓存创建

5m: 0 / 1h: 35

5m: 0 / 1h: 43

5m: 0 / 1h: 41

5m: 0 / 1h: 35

缓存读取 32,231 32,545 0 32,562 32,564

输出速度 - 41.3 tok/s 43.1 tok/s 58.6 tok/s 114.6 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。

我是 Claude，由 Anthropic 开发的 AI 助手。我基于大型语言模型构建，能够帮助用户完成写作、分析、编程、问答等各种任务。

我是 Kiro，一个由 Anthropic 技术驱动的 AI 开发环境，基于 Claude Sonnet 4.6 模型构建，专注于帮助开发者完成软件工程任务。

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

我是 Claude，由 Anthropic 开发的 AI 助手。目前运行的模型版本是 claude-sonnet-4-6。

步骤 5: knowledge_recall text

时间 12:42:51 ~ 12:42:56 (4.9s) 12:46:55 ~ 12:47:05 (5.6s) 12:47:18 ~ 12:47:28 (5.1s) 12:51:19 ~ 12:51:29 (5.3s) 12:49:59 ~ 12:50:09 (9.0s)

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) 2,756 ms 2,226 ms 2,481 ms 2,031 ms 5,774 ms

首文本(SSE delta) 3,833 ms 2,228 ms 2,642 ms 2,032 ms 5,776 ms

总耗时 3,867 ms 4,886 ms 4,294 ms 4,362 ms 7,807 ms

输入 3 3 0 3 3

输出 198 205 221 205 205

缓存创建

7,679

5m: 0 / 1h: 7,679

8,039

5m: 0 / 1h: 8,039

8,043

5m: 0 / 1h: 8,043

8,032

5m: 0 / 1h: 8,032

缓存读取 0 0 0 0 0

输出速度 burst 77.1 tok/s 133.8 tok/s 88.0 tok/s 100.9 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 12:45:23 ~ 12:45:38 (14.8s) 12:49:38 ~ 12:49:58 (19.5s) 12:50:36 ~ 12:50:56 (15.3s) 12:52:42 ~ 12:53:13 (26.0s) 12:53:53 ~ 12:54:18 (23.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=694311116986036897795628997160874806262549390996772679404983223262307839

HTTP 头(TTFB) 1,655 ms 2,156 ms 2,128 ms 1,349 ms 4,141 ms

首文本(SSE delta) 2,074 ms 2,557 ms 2,143 ms 1,372 ms 4,143 ms

总耗时 13,771 ms 18,708 ms 14,658 ms 24,983 ms 22,768 ms

输入 110 164 0 164 164

输出 1,062 1,289 991 1,975 1,768

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 90.8 tok/s 79.8 tok/s 79.2 tok/s 83.6 tok/s 94.9 tok/s

响应

```json
{"n":72,"c":[6,8,13]}
```

**Verification breakdown:**

String: `694311116986036897795628997160874806262549390996772679404983223262307839`

- **Length (n):** 72 digits
- **0s (count0):** posit

## Counting digits in s

**s** = `694311116986036897795628997160874806262549390996772679404983223262307839`

### Length
Splitting into groups of 10:
`6943111169` | `8603689779` | `5628997160` | `87480

```json
{"n":72,"c":[6,8,13]}
```

## Counting Process

**String:**
`694311116986036897795628997160874806262549390996772679404983223262307839`

Indexing each character:

| Pos | Char | Pos | Char | Pos | Char | Pos | Char |
|-----|----

## Counting digits in s

**s** = `694311116986036897795628997160874806262549390996772679404983223262307839`

### Step-by-step indexing (1-based):

| # | digit | # | digit | # | digit | # | digit |
|--

总计

总输入(含缓存) 136,460 138,176 0 138,212 138,199

总输出 1,319 1,591 1,274 2,266 2,052

总缓存创建

39,945

5m: 0 / 1h: 39,945

40,627

5m: 0 / 1h: 40,627

40,646

5m: 0 / 1h: 40,646

40,631

5m: 0 / 1h: 40,631

总缓存读取 96,390 97,370 0 97,387 97,389

输出速度 90.8 tok/s 76.6 tok/s 83.0 tok/s 83.2 tok/s 95.9 tok/s

总执行时间 36.8s 45.4s 36.6s 49.6s 1m 1.2s

总墙钟时间

8m 40.7s

含步骤间等待

13m 3.0s

含步骤间等待

13m 55.0s

含步骤间等待

16m 13.2s

含步骤间等待

17m 14.6s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1989

Anthropic: Claude Sonnet 4.6

$0.2060

Anthropic: Claude Sonnet 4.6

>=$0.0191

Anthropic: Claude Sonnet 4.6 (成本不完整)

$0.2162

Anthropic: Claude Sonnet 4.6

$0.2129

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

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

CDN cloudflare cloudflare - - -

ID 格式 msg_01QkNw1hhsUPcVHTCheeBuy9 msg_01VLQP6nHeJerz5LE6jNaJGZ msg_14db80c74c62470f81dc180182a4e2b2 msg_014ZhdyeMpGtU61Wr7ADxGGs msg_012t2xbjq2coFhLkZiEL4DtN

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-Max	cf-ray: a12a0ab4eacfb35c-AMS server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606280445246607594608268d9d6W3XyG4IG	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	a12a0ab4eacfb35c-AMS
M-	cf-ray: a12a10ee3aa2e393-NRT server: cloudflare set-cookie: server_name_session=3de609b8cf2a7e28d141c3b6c8b41f57; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 20260628044940209895692gIpusbEN	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	a12a10ee3aa2e393-NRT
R-Kiro	server: Photon-Edge x-amzn-requestid: 2129a7a5-f3a6-423d-8e30-b55df4aecc82	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	1782622237704-y5a9mmguo96zv 2129a7a5-f3a6-423d-8e30-b55df4aecc82
M-Max	server: nginx x-new-api-version: 0b8be5cf x-oneapi-request-id: 2026062812524485116309jGU3QmUi	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	—
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_011CcV4TELppEShpndKE2RRt

协议指纹评分（机器自动）

总分 / 100 86 98 50 98 100

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

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

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

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

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

系统提示纯净 5.0 /10 ×8 10.0 /10 ×8 未计入 10.0 /10 ×8 10.0 /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

知识截止 5.0 /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 0.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

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

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

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

流式投递 0.0 /10 ×5 10.0 /10 ×5 10.0 /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 10.0 /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 86 v3.25.0 98 v3.25.0 50 v3.25.0 98 v3.25.0 基准（参考）

雷达图例

共 20 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 档位思考量 tier_thinking_volume_match
20 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 5036 ms 6634 ms 5004 ms 7238 ms 9128 ms

输出速度 90.8 tok/s 76.6 tok/s 83.0 tok/s 83.2 tok/s 95.9 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 ⓘ

$0.1989

$0.2060

>=$0.0191 (成本不完整)

$0.2162

$0.2129

缓存读取占比 ⓘ 71% 70% - 70% 70%

揭晓

通道

服务商 AIMZ ↗

通道M-Max

服务商 LingxiCode ↗

通道M-

服务商 SSSAiCode ↗

通道R-Kiro

服务商 SudoCode ↗

通道M-Max

官方基准

返回列表