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

2026/06/21 运行日期

claude-sonnet-4-6 官方基准

1 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

步骤 / 指标

Novaris

O-Max

TopRouterCN

O-Max

官方基准

DawAPI

SSSAiCode

O-Max-L2

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 04:37:12 ~ 04:37:22 (5.0s) 04:37:14 ~ 04:37:24 (5.1s) 04:37:22 ~ 04:37:27 (4.3s) 04:37:10 ~ 04:37:20 (5.4s) 04:37:20 ~ 04:37:25 (3.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,787 ms 3,973 ms 3,007 ms 3,915 ms 2,304 ms

首文本(SSE delta) 3,838 ms 3,975 ms 3,009 ms 3,917 ms 2,388 ms

总耗时 3,954 ms 4,050 ms 3,114 ms 4,089 ms 2,463 ms

输入 3 1,467 3 3 3

输出 5 5 5 5 5

缓存创建

25,697

5m: 25,697 / 1h: 0

30,854

5m: 0 / 1h: 30,854

32,318

5m: 0 / 1h: 32,318

32,346

5m: 32,346 / 1h: 0

31,990

5m: 0 / 1h: 31,990

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 04:39:17 ~ 04:39:22 (3.2s) 04:40:37 ~ 04:40:42 (4.2s) 04:40:35 ~ 04:40:40 (4.2s) 04:38:50 ~ 04:39:00 (5.5s) 04:41:16 ~ 04:41:21 (3.4s)

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。严格按以下 3 行格式输出，全部小写英文，不要任何其他文字、解释、标点：
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) 1,864 ms 3,188 ms 2,963 ms 4,294 ms 2,153 ms

首文本(SSE delta) 1,865 ms 3,189 ms 2,964 ms 4,486 ms 2,444 ms

总耗时 2,402 ms 3,300 ms 3,173 ms 4,716 ms 2,521 ms

输入 3 3 3 3 3

输出 22 22 22 22 22

缓存创建

191

5m: 191 / 1h: 0

1,655

5m: 0 / 1h: 1,655

191

5m: 0 / 1h: 191

191

5m: 191 / 1h: 0

191

5m: 0 / 1h: 191

缓存读取 25,697 30,854 32,318 32,346 31,990

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-4-5

步骤 3: cutoff text

时间 04:41:39 ~ 04:41:44 (3.9s) 04:43:42 ~ 04:43:52 (5.2s) 04:42:05 ~ 04:42:10 (4.4s) 04:40:43 ~ 04:40:53 (5.0s) 04:44:18 ~ 04:44:23 (3.8s)

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) 2,626 ms 4,195 ms 3,112 ms 3,970 ms 2,516 ms

首文本(SSE delta) 2,829 ms 4,196 ms 3,114 ms 3,971 ms 2,643 ms

总耗时 3,149 ms 4,287 ms 3,354 ms 4,176 ms 2,717 ms

输入 3 3 3 3 3

输出 8 8 8 8 8

缓存创建

5m: 55 / 1h: 0

5m: 0 / 1h: 38

5m: 0 / 1h: 55

5m: 55 / 1h: 0

5m: 0 / 1h: 57

缓存读取 25,888 32,509 32,509 32,537 32,181

响应

2025-04

2025-08

2025-08

2025-08

2025-04

步骤 4: identity_free text

时间 04:43:31 ~ 04:43:36 (4.4s) 04:47:13 ~ 04:47:23 (5.7s) 04:44:37 ~ 04:44:42 (4.5s) 04:43:45 ~ 04:43:50 (4.0s) 04:46:49 ~ 04:46:54 (3.5s)

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) 2,544 ms 4,827 ms 2,790 ms 2,319 ms 1,890 ms

首文本(SSE delta) 2,546 ms 4,828 ms 2,792 ms 2,730 ms 2,744 ms

总耗时 3,631 ms 4,899 ms 3,496 ms 3,030 ms 2,812 ms

输入 3 3 3 3 3

输出 66 64 50 50 57

缓存创建

5m: 43 / 1h: 0

5m: 0 / 1h: 43

5m: 43 / 1h: 0

5m: 0 / 1h: 43

缓存读取 25,943 32,547 32,564 32,592 32,238

输出速度 60.8 tok/s - 71.0 tok/s - -

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助回答问题、分析信息、编写代码、创作内容等，致力于提供安全、有益的对话体验。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助完成写作、分析、编程、问答等多种任务，致力于提供安全、有益的对话体验。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助回答问题、分析代码、撰写文本等多种任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我可以帮助回答问题、分析代码、写作和解决各类任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助编程、写作、分析和回答问题，致力于提供安全、有用的回应。

步骤 5: knowledge_recall text

时间 04:45:05 ~ 04:45:15 (6.4s) 04:51:20 ~ 04:51:30 (5.1s) 04:46:33 ~ 04:46:43 (5.0s) 04:45:22 ~ 04:45:32 (5.5s) 04:50:46 ~ 04:50:56 (5.4s)

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) 1,527 ms 4,156 ms 2,130 ms 2,715 ms 2,295 ms

首文本(SSE delta) 1,920 ms 4,157 ms 2,132 ms 2,756 ms 4,455 ms

总耗时 5,639 ms 4,203 ms 4,072 ms 4,697 ms 4,517 ms

输入 3 3 3 3 3

输出 446 205 205 205 198

缓存创建

2,360

5m: 2,360 / 1h: 0

8,043

5m: 0 / 1h: 8,043

8,046

5m: 0 / 1h: 8,046

8,074

5m: 8,074 / 1h: 0

7,727

5m: 0 / 1h: 7,727

缓存读取 0 0 0 0 0

输出速度 119.9 tok/s burst 105.7 tok/s 105.6 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 04:48:27 ~ 04:48:47 (16.6s) 04:53:01 ~ 04:53:26 (21.0s) 04:50:27 ~ 04:50:47 (16.7s) 04:49:15 ~ 04:49:50 (30.0s) 04:53:33 ~ 04:53:38 (1.1s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=886482659178159229794163591468450442227160175565088943304196784355022684

HTTP 头(TTFB) 1,689 ms 6,140 ms 1,033 ms 1,564 ms -

首文本(SSE delta) 1,689 ms 6,141 ms 1,035 ms 26,674 ms -

总耗时 15,774 ms 20,200 ms 15,677 ms 29,218 ms -

输入 154 164 164 193 -

输出 1,282 1,881 1,416 2,569 -

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 -

输出速度 91.0 tok/s 133.8 tok/s 96.7 tok/s burst -

响应

## Counting digits in s

**s:** `886482659178159229794163591468450442227160175565088943304196784355022684`

### Step 1 — Split into groups of 10 for accuracy

```
Position  | Chunk
----------|--------

## Counting digits in s

**s = `886482659178159229794163591468450442227160175565088943304196784355022684`**

### Step 1 — Length

Breaking into groups of 10:
```
8864826591  (10)
7815922979  (10)
4163

```
s = "886482659178159229794163591468450442227160175565088943304196784355022684"
```

Indexing each character:

| pos | 1–10 | 11–20 | 21–30 | 31–40 | 41–50 | 51–60 | 61–70 | 71–72 |
|-----|------|-

```python
s = "886482659178159229794163591468450442227160175565088943304196784355022684"
n = len(s)
c0 = s.count('0')
c7 = s.count('7')
c9 = s.count('9')
```

**Verification by groups of 10:**

| Grou

失败

评测失败

总计

总输入(含缓存)

106,043

138,186

138,223

138,392

136,432

仅前 6 步

总输出

1,829

2,185

1,706

2,859

290

仅前 6 步

总缓存创建

28,346

5m: 28,346 / 1h: 0

40,633

5m: 0 / 1h: 40,633

40,653

5m: 0 / 1h: 40,653

40,709

5m: 40,709 / 1h: 0

40,008

5m: 0 / 1h: 40,008

仅前 6 步

总缓存读取

77,528

95,910

97,391

97,475

96,409

仅前 6 步

输出速度

95.0 tok/s

133.8 tok/s

96.7 tok/s

105.6 tok/s

仅前 6 步

总执行时间

39.4s

46.2s

39.1s

55.5s

20.9s

仅前 6 步

总墙钟时间

11m 35.0s

含步骤间等待

16m 12.5s

含步骤间等待

13m 25.0s

含步骤间等待

12m 40.0s

含步骤间等待

16m 17.9s

含步骤间等待

仅前 6 步

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1176

Anthropic: Claude Sonnet 4.6

$0.2189

Anthropic: Claude Sonnet 4.6

$0.2078

Anthropic: Claude Sonnet 4.6

$0.1682

Anthropic: Claude Sonnet 4.6

$0.1833

Anthropic: Claude Sonnet 4.6 (部分)

通道指纹

来源步骤 digit_count digit_count digit_count digit_count knowledge_recall

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

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

CDN - - - - -

ID 格式 msg_018ubqJTYYxhk9n4oRqaCf4x msg_0137dyu7myx6VmFFzfsWF3M6 msg_0123r5CdKxwgCbcKT2VXuBCG msg_01HHPxHryw1uVupavc4pmjjW msg_01SBeP21tAbq1wiKsC75Z2F1

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	server: nginx x-oneapi-request-id: 202606202048281913750458268d9d63p5ST3Sn	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-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
O-Max	server: nginx/1.27.5	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	tr-ew7y4grd-bnfxa1qgw8-112118
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_011CcFAvXGTPaXTaaKxs8LCm
cc	server: cdn x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606202049161586250458268d9d6ZgZPC8cX	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	—
O-Max-L2	cf-ray: a0eda5d96ae4f7d1-LAX server: Photon-Edge set-cookie: _cfuvid=16pjem45SpPsmA__O7RQGZkQndr7YhMwRU37yqRh6sM-1781988647.906413-1.0.1.1-BRzkLhItLPV9f5wCs.TsugHBnFxNMhlfzpGxNIMZwtY; 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	1781988647864-8m6a1l68jczxh req_011CcFAwwtKVQEHyXULDPrKF a0eda5d96ae4f7d1-LAX

协议指纹评分（机器自动）

总分 / 100 75 95 100 89

基线状态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 —

缓存 TTL 一致 0.0 /10 ×15 10.0 /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 —

缓存连续性 10.0 /10 ×14 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 10.0 /10 ×13 —

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

系统提示纯净 0.0 /10 ×8 10.0 /10 ×8 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 —

知识截止 5.0 /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 —

service_tier 10.0 /10 ×6 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 10.0 /10 ×5 —

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

流式投递 0.0 /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 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 —

SDK 一致 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 75 v3.24.0 95 v3.24.0 基准（参考） 89 v3.24.0 执行失败，不评分

雷达图例

共 22 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 Thinking 存在 thinking_present
5 流式投递 buffer_dump_match
6 自由身份 identity_free_clean
7 停止原因 stop_reason_present
8 系统提示纯净 system_prompt_clean
9 Service Tier service_tier_present
10 缓存命中比 cache_hit_ratio_match
11 缓存 TTL 一致 cache_ttl_consistency
12 推理地区 inference_geo_present
13 缓存连续 cache_continuity_intra
14 知识召回 knowledge_recall_match
15 延迟基线 latency_baseline_match
16 原生 msg-ID anthropic_msg_id_format
17 Sliding 缓存 cache_sliding_correctness
18 结构化身份 identity_structured_match
19 信封自报一致 envelope_self_report_match
20 中文遵循 instruction_following_lang
21 档位思考量 tier_thinking_volume_match
22 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 5758 ms 6823 ms 5481 ms 8321 ms 3006 ms

输出速度 95.0 tok/s 133.8 tok/s 96.7 tok/s 105.6 tok/s -

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

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1176

$0.2189

$0.2078

$0.1682

$0.1833 (部分)

缓存读取占比 ⓘ 73% 69% 70% 70% 71% (部分)

揭晓

通道

服务商 Novaris ↗

通道O-Max

服务商 TopRouterCN ↗

通道O-Max

官方基准

服务商 DawAPI ↗

通道cc

服务商 SSSAiCode ↗

通道O-Max-L2

返回列表