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

2026/06/24 运行日期

claude-sonnet-4-6 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

AiTokensFlux

O-Pro

AnyAIGC

M-

LinkAPI

O-Max

官方基准

FoxCode

R-ultra

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 05:36:29 ~ 05:36:39 (5.6s) 05:36:35 ~ 05:36:40 (4.3s) 05:36:31 ~ 05:36:41 (5.1s) 05:36:37 ~ 05:36:47 (5.2s) 05:36:33 ~ 05:36:38 (4.5s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 4,194 ms 2,801 ms 3,448 ms 3,776 ms 2,603 ms

首文本(SSE delta) 4,196 ms 2,802 ms 3,449 ms 3,778 ms 2,604 ms

总耗时 4,703 ms 2,830 ms 3,667 ms 3,978 ms 3,380 ms

输入 3 8,598 5,934 3 71

输出 5 2 5 5 14

缓存创建

32,457

5m: 0 / 1h: 32,457

25,697

5m: 25,697 / 1h: 0

32,318

5m: 0 / 1h: 32,318

32,947

5m: 32,947 / 1h: 0

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 05:38:53 ~ 05:38:58 (4.8s) 05:39:11 ~ 05:39:16 (4.7s) 05:39:28 ~ 05:39:38 (7.5s) 05:39:41 ~ 05:39:46 (3.6s) 05:40:00 ~ 05:40:10 (5.1s)

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) 3,030 ms 3,449 ms 6,179 ms 2,266 ms 3,351 ms

首文本(SSE delta) 3,031 ms 3,450 ms 6,203 ms 2,268 ms 3,352 ms

总耗时 3,492 ms 3,560 ms 6,544 ms 2,623 ms 4,214 ms

输入 3 8,797 3 3 24

输出 22 21 22 22 79

缓存创建

192

5m: 0 / 1h: 192

6,123

5m: 6,123 / 1h: 0

192

5m: 0 / 1h: 192

33,512

5m: 33,512 / 1h: 0

缓存读取 32,457 0 25,697 32,318 0

输出速度 - - - - 91.6 tok/s

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 05:41:11 ~ 05:41:16 (4.1s) 05:41:06 ~ 05:41:11 (4.1s) 05:42:44 ~ 05:42:59 (11.7s) 05:43:23 ~ 05:43:28 (4.4s) 05:42:06 ~ 05:42:16 (6.6s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,033 ms 2,755 ms 10,527 ms 2,766 ms 5,183 ms

首文本(SSE delta) 3,160 ms 2,812 ms 10,684 ms 2,768 ms 5,184 ms

总耗时 3,228 ms 2,888 ms 10,699 ms 3,427 ms 5,701 ms

输入 3 8,836 3 3 68

输出 8 5 8 8 50

缓存创建

5m: 0 / 1h: 38

5m: 57 / 1h: 0

5m: 0 / 1h: 38

370

5m: 370 / 1h: 0

缓存读取 32,649 0 31,820 32,510 33,512

输出速度 - - - - 96.7 tok/s

响应

2025-08

2024-05

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 05:45:04 ~ 05:45:14 (6.2s) 05:44:50 ~ 05:45:00 (5.4s) 05:45:46 ~ 05:45:56 (7.3s) 05:45:57 ~ 05:46:02 (4.8s) 05:43:38 ~ 05:43:49 (5.4s)

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) 4,293 ms 4,005 ms 5,192 ms 3,204 ms 3,254 ms

首文本(SSE delta) 4,294 ms 4,006 ms 5,213 ms 3,206 ms 3,255 ms

总耗时 5,328 ms 4,231 ms 6,236 ms 3,773 ms 4,159 ms

输入 3 8,877 3 3 68

输出 67 25 60 49 70

缓存创建

5m: 0 / 1h: 43

5m: 43 / 1h: 0

5m: 0 / 1h: 43

34,283

5m: 34,283 / 1h: 0

缓存读取 32,687 0 31,877 32,548 0

输出速度 64.8 tok/s - 58.7 tok/s 86.4 tok/s 77.4 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助你完成写作、分析、编程、问答等各类任务，致力于提供安全、有益的智能对话服务。

我是Anthropic开发的Claude智能助理，专注于安全、高效的人工智能对话与辅助。

我是 Claude，由 Anthropic 开发的 AI 助手。当前运行的模型为 claude-sonnet-4-6，擅长对话、分析、编程和写作等任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助你完成写作、分析、编程、问答等各种任务。

我是 Kiro，一个由 Anthropic 开发的 AI 编程助手，基于 Claude 模型构建，帮助开发者完成软件工程任务。

步骤 5: knowledge_recall text

时间 05:47:30 ~ 05:47:40 (5.5s) 05:48:47 ~ 05:48:57 (7.3s) 05:47:48 ~ 05:47:58 (5.8s) 05:48:05 ~ 05:48:15 (6.2s) 05:45:08 ~ 05:45:38 (27.6s)

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) 2,533 ms 4,360 ms 3,049 ms 2,898 ms 2,671 ms

首文本(SSE delta) 2,534 ms 4,361 ms 3,050 ms 2,899 ms 2,672 ms

总耗时 4,678 ms 6,080 ms 4,868 ms 5,239 ms 26,337 ms

输入 3 9,504 3 3 21

输出 204 154 198 204 2,232

缓存创建

8,047

5m: 0 / 1h: 8,047

8,085

5m: 8,085 / 1h: 0

8,029

5m: 0 / 1h: 8,029

9,706

5m: 9,706 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 95.1 tok/s 89.6 tok/s 108.9 tok/s 87.2 tok/s 94.3 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":"UNKNOWN","oscars2024_best_picture":"UNKNOWN","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_peace_2024":"UNKNOWN","wo

{"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 05:48:53 ~ 05:49:13 (17.0s) 05:50:47 ~ 05:50:52 (2.7s) 05:50:56 ~ 05:51:16 (17.2s) 05:51:42 ~ 05:52:07 (21.6s) 05:49:20 ~ 05:49:35 (13.3s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=719745329626877435149580268412972553658732082180100065224063071513982232

HTTP 头(TTFB) 1,407 ms 1,456 ms 1,347 ms 6,710 ms 1,975 ms

首文本(SSE delta) 1,408 ms 1,457 ms 1,348 ms 6,712 ms 1,976 ms

总耗时 16,092 ms 1,570 ms 16,250 ms 20,575 ms 12,537 ms

输入 164 166 164 164 69

输出 1,363 18 1,322 1,345 1,187

缓存创建

5m: 0 / 1h: 0

140

5m: 140 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 92.8 tok/s - 88.7 tok/s 97.0 tok/s 112.4 tok/s

响应

## Counting digits in `s`

**String:** `719745329626877435149580268412972553658732082180100065224063071513982232`

### Length
Splitting into groups of 10:
```
7197453296 | 2687743514 | 9580268412 | 97

```json
{"n":68,"c":[5,6,7]}
```

**Step 1 – Index every character (1-based)**

```
 1:7  2:1  3:9  4:7  5:4  6:5  7:3  8:2  9:9 10:6
11:2 12:6 13:8 14:7 15:7 16:4 17:3 18:5 19:1 20:4
21:9 22:5 23:8 24:0 25:2 26:6 27:8 28:4 29:1 30:2

```json
{"n":72,"c":[8,7,5]}
```

**Verification:**

`719745329626877435149580268412972553658732082180100065224063071513982232`

| | |
|---|---|
| **Length** | 72 |
| **0s** (positions 24,43,48,50,51,

{"n":72,"c":[8,7,5]}

总计

总输入(含缓存) 138,749 44,778 135,509 138,175 144,791

总输出 1,669 225 1,615 1,633 3,632

总缓存创建

40,777

5m: 0 / 1h: 40,777

40,005

5m: 40,005 / 1h: 0

40,620

5m: 0 / 1h: 40,620

110,958

5m: 110,958 / 1h: 0

总缓存读取 97,793 0 89,394 97,376 33,512

输出速度 91.5 tok/s 89.6 tok/s 89.0 tok/s 95.3 tok/s 99.1 tok/s

总执行时间 43.2s 28.6s 54.5s 45.8s 1m 2.6s

总墙钟时间

12m 44.0s

含步骤间等待

14m 16.9s

含步骤间等待

14m 45.1s

含步骤间等待

15m 29.9s

含步骤间等待

13m 2.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2078

Anthropic: Claude Sonnet 4.6

$0.1377

Anthropic: Claude Sonnet 4.6

$0.1631

Anthropic: Claude Sonnet 4.6

$0.2066

Anthropic: Claude Sonnet 4.6

$0.3256

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

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

CDN - - - - cloudflare

ID 格式 msg_01Mx8ig8H1ahqDx7uiZ83Ayy 自定义 msg_01PQzdR1gKVs5vNPz9dfy13U msg_01M2yGHdHPWsFPrHNvC13KoF msg_017QwBS0POdlV3OrIJ7USats

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-Pro	server: nginx	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	—
M-	server: nginx	id_format: non_standard has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	body:id:non_standardbody:usage.cache_fields	—
O-Max	server: nginx set-cookie: server_name_session=7d88fca8b1426ca76c58d5aeba6fb224; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606232150574779069488268d9d6LoXoceAl	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_011CcLw29MtXwUpgMDQg4WH6
R-ultra	cf-ray: a106b3bf8ad47963-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606232149213655780058268d9d6bQ0XPV8n	id_format: msg_prefix 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_fieldsupstream:anthropic-direct-from-body	a106b3bf8ad47963-NRT

协议指纹评分（机器自动）

总分 / 100 97 37 84 100 68

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

缓存 TTL 一致 10.0 /10 ×15 0.0 /10 ×15 0.0 /10 ×15 10.0 /10 ×15 0.0 /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 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 5.3 /10 ×13

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

系统提示纯净 10.0 /10 ×8 0.0 /10 ×8 5.0 /10 ×8 10.0 /10 ×8 5.0 /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 0.0 /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 0.0 /10 ×7

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

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

延迟基线 10.0 /10 ×5 10.0 /10 ×5 3.1 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

流式投递 10.0 /10 ×5 3.9 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 6.6 /10 ×5

Req-ID 透传 0.0 /10 ×4 0.0 /10 ×4 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 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 97 v3.24.0 37 v3.24.0 84 v3.24.0 基准（参考） 68 v3.24.0

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 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 6254 ms 3527 ms 8044 ms 6603 ms 9388 ms

输出速度 91.5 tok/s 89.6 tok/s 89.0 tok/s 95.3 tok/s 99.1 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 ⓘ

$0.2078

$0.1377

$0.1631

$0.2066

$0.3256

缓存读取占比 ⓘ 70% 0% 66% 70% 23%

揭晓

通道

服务商 AiTokensFlux ↗

通道O-Pro

服务商 AnyAIGC ↗

通道M-

服务商 LinkAPI ↗

通道O-Max

官方基准

服务商 FoxCode ↗

通道R-ultra

返回列表