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

2026/06/18 运行日期

claude-sonnet-4-6 官方基准

全部 ≥ 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

duckcoding

cc专用

ClaudeAPI

O-Bdrk

官方基准

SudoCode

M-Max

OmniaKey

O-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 23:46:04 ~ 23:46:09 (4.4s) 23:46:06 ~ 23:46:16 (7.3s) 23:46:10 ~ 23:46:15 (4.1s) 23:46:08 ~ 23:46:18 (5.7s) 23:46:02 ~ 23:46:12 (6.2s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,103 ms 2,777 ms 2,850 ms 4,089 ms 4,607 ms

首文本(SSE delta) 3,105 ms 2,779 ms 2,852 ms 4,105 ms 4,608 ms

总耗时 3,313 ms 3,139 ms 2,999 ms 4,592 ms 4,809 ms

输入 3 3 3 3 3

输出 5 5 5 5 5

缓存创建

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

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 23:48:27 ~ 23:48:32 (4.0s) 23:47:27 ~ 23:47:37 (5.7s) 23:49:20 ~ 23:49:25 (3.3s) 23:48:15 ~ 23:48:25 (5.0s) 23:48:36 ~ 23:48:46 (5.0s)

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。请只输出以下 3 行内容（label 小写，值小写）：
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) 2,668 ms 2,476 ms 2,026 ms 3,301 ms 3,335 ms

首文本(SSE delta) 2,669 ms 2,478 ms 2,027 ms 3,302 ms 3,336 ms

总耗时 3,327 ms 3,635 ms 2,369 ms 3,673 ms 3,678 ms

输入 3 3 3 3 3

输出 22 22 22 22 22

缓存创建

178

5m: 0 / 1h: 178

178

5m: 0 / 1h: 178

178

5m: 0 / 1h: 178

178

5m: 0 / 1h: 178

178

5m: 0 / 1h: 178

缓存读取 32,318 32,318 32,318 32,318 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 23:50:41 ~ 23:50:46 (4.3s) 23:51:06 ~ 23:51:16 (5.9s) 23:50:29 ~ 23:50:34 (3.3s) 23:50:17 ~ 23:50:22 (4.8s) 23:50:57 ~ 23:51:07 (7.0s)

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) 3,139 ms 3,191 ms 2,083 ms 3,477 ms 5,307 ms

首文本(SSE delta) 3,256 ms 3,192 ms 2,085 ms 3,479 ms 5,308 ms

总耗时 3,486 ms 3,568 ms 2,292 ms 3,657 ms 6,023 ms

输入 3 3 3 3 3

输出 8 8 8 8 8

缓存创建

5m: 0 / 1h: 57

5m: 0 / 1h: 55

5m: 0 / 1h: 57

32,553

5m: 0 / 1h: 32,553

5m: 0 / 1h: 55

缓存读取 32,496 32,496 32,496 0 32,496

响应

2025-08

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 23:53:55 ~ 23:54:00 (4.8s) 23:53:53 ~ 23:54:03 (6.5s) 23:52:26 ~ 23:52:31 (4.2s) 23:54:08 ~ 23:54:19 (5.0s) 23:53:35 ~ 23:53:45 (7.7s)

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) 3,445 ms 3,704 ms 2,358 ms 3,122 ms 3,547 ms

首文本(SSE delta) 3,446 ms 3,706 ms 2,359 ms 3,123 ms 3,548 ms

总耗时 3,843 ms 5,101 ms 3,046 ms 3,930 ms 4,392 ms

输入 3 3 3 3 3

输出 24 66 49 39 55

缓存创建

5m: 0 / 1h: 35

5m: 0 / 1h: 43

5m: 0 / 1h: 39

5m: 0 / 1h: 43

缓存读取 32,553 32,551 32,553 32,553 32,551

输出速度 - 47.3 tok/s 71.3 tok/s 48.3 tok/s 65.2 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、有帮助的对话体验。

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

我是 Claude，由 Anthropic 打造的 AI 助手。当前版本为 claude-sonnet-4-6。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助您解答问题、编写代码、分析文本、创作内容等多种任务。

步骤 5: knowledge_recall text

时间 23:57:22 ~ 23:57:32 (5.9s) 23:56:17 ~ 23:56:27 (7.5s) 23:54:11 ~ 23:54:21 (5.5s) 23:55:28 ~ 23:55:38 (6.1s) 23:55:33 ~ 23:55:43 (7.3s)

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) 1,572 ms 2,037 ms 1,713 ms 2,612 ms 3,060 ms

首文本(SSE delta) 1,591 ms 2,039 ms 1,716 ms 2,614 ms 3,062 ms

总耗时 3,556 ms 4,780 ms 4,096 ms 5,012 ms 6,020 ms

输入 3 3 3 3 3

输出 205 205 205 204 205

缓存创建

8,001

5m: 0 / 1h: 8,001

8,049

5m: 0 / 1h: 8,049

8,034

5m: 0 / 1h: 8,034

8,020

5m: 0 / 1h: 8,020

8,038

5m: 0 / 1h: 8,038

缓存读取 0 0 0 0 0

输出速度 104.3 tok/s 74.8 tok/s 86.1 tok/s 85.1 tok/s 69.3 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":"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 23:59:01 ~ 23:59:16 (11.5s) 23:59:29 ~ 23:59:49 (18.5s) 23:57:20 ~ 23:57:40 (17.0s) 23:58:27 ~ 23:58:42 (11.5s) 23:57:32 ~ 23:57:57 (21.5s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=114285708408588284689028408000528291308542927609292063843815949552101188

HTTP 头(TTFB) 1,041 ms 1,537 ms 1,434 ms 1,706 ms 1,630 ms

首文本(SSE delta) 1,042 ms 1,538 ms 1,813 ms 1,719 ms 2,269 ms

总耗时 10,316 ms 16,995 ms 16,008 ms 10,383 ms 19,163 ms

输入 164 164 164 164 164

输出 881 1,298 1,128 710 1,389

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 95.0 tok/s 84.0 tok/s 79.5 tok/s 81.9 tok/s 82.2 tok/s

响应

```json
{"n":72,"c":[11,2,7]}
```

## Counting digits in:
`114285708408588284689028408000528291308542927609292063843815949552101188`

### Length
Grouping by 10s:
```
1142857084  (10)
0858828468  (10)
9028408000  (10)
5282913085  (10)
4

## Counting digits in s

**s** = `114285708408588284689028408000528291308542927609292063843815949552101188`

### Step-by-step breakdown (groups of 10):

| Pos | Chars | 0s | 7s | 9s |
|-----|-------|-

```json
{"n":72,"c":[11,2,7]}
```

## Counting digits in s

**s** = `114285708408588284689028408000528291308542927609292063843815949552101188`

### Length count (groups of 10):
```
1142857084  → 10
0858828468  → 10
9028408000  → 10
528

总计

总输入(含缓存) 138,135 138,187 138,176 138,158 138,176

总输出 1,145 1,604 1,417 988 1,684

总缓存创建

40,589

5m: 0 / 1h: 40,589

40,643

5m: 0 / 1h: 40,643

40,630

5m: 0 / 1h: 40,630

73,108

5m: 0 / 1h: 73,108

40,632

5m: 0 / 1h: 40,632

总缓存读取 97,367 97,365 97,367 64,871 97,365

输出速度 96.6 tok/s 80.1 tok/s 80.1 tok/s 80.3 tok/s 79.7 tok/s

总执行时间 34.9s 51.3s 37.4s 38.0s 54.9s

总墙钟时间

13m 12.0s

含步骤间等待

13m 43.0s

含步骤间等待

11m 29.9s

含步骤间等待

12m 34.0s

含步骤间等待

11m 55.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1991

Anthropic: Claude Sonnet 4.6

$0.2062

Anthropic: Claude Sonnet 4.6

$0.2034

Anthropic: Claude Sonnet 4.6

$0.3090

Anthropic: Claude Sonnet 4.6

$0.2074

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 (100%) one-api (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_01CFyZ6teBFfnw6AjevcPa3e msg_01YahyZLLBLKhVoAyxBJWCSf msg_01QGx2xTk8Ts2VKL1ABRxPde msg_01V4pGqKgBNx4qovKrQKtWXq msg_01DdQLm9JXwdJ69SS8pTL5v6

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
cc专用	server: Nginx x-oneapi-request-id: 202606181559026040918338268d9d6IUPnycKj	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-Bdrk	x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606181559309387380618268d9d6zd87wI00	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_011CcAzxEjfFgkgk86wd4Crx
M-Max	server: nginx x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606182358285029183254YPJU1A1	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	cf-ray: a0db7d946c12f7ea-AMS server: cloudflare x-oneapi-request-id: 202606181557348604516958268d9d6QQ77ky1T	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	a0db7d946c12f7ea-AMS

协议指纹评分（机器自动）

总分 / 100 98 98 100 90 98

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 10.0 /10 ×20 5.5 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 10.0 /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 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 7.7 /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 10.0 /10 ×8

系统提示纯净 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 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 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 10.0 /10 ×6

inference_geo 10.0 /10 ×5 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 10.0 /10 ×5

流式投递 10.0 /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 10.0 /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 98 v3.24.0 基准（参考） 90 v3.24.0 98 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 4640 ms 6203 ms 5135 ms 5208 ms 7348 ms

输出速度 96.6 tok/s 80.1 tok/s 80.1 tok/s 80.3 tok/s 79.7 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 ⓘ

$0.1991

$0.2062

$0.2034

$0.3090

$0.2074

缓存读取占比 ⓘ 70% 70% 70% 47% 70%

揭晓

通道

服务商 duckcoding ↗

通道cc专用

服务商 ClaudeAPI ↗

通道O-Bdrk

官方基准

服务商 SudoCode ↗

通道M-Max

服务商 OmniaKey ↗

通道O-Max

返回列表