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

2026/06/27 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

OmniaKey

O-Max

官方基准

duckcoding

cc专用

TopRouterCN

O-Max

PackyCode

cc-Zeus

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 21:34:23 ~ 21:34:33 (6.3s) 21:34:31 ~ 21:34:36 (3.5s) 21:34:29 ~ 21:34:39 (6.4s) 21:34:25 ~ 21:34:30 (4.1s) 21:34:27 ~ 21:34:32 (3.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 4,654 ms 2,233 ms 3,393 ms 3,219 ms 2,218 ms

首文本(SSE delta) 4,655 ms 2,258 ms 3,523 ms 3,221 ms 2,219 ms

总耗时 4,774 ms 2,365 ms 3,660 ms 3,288 ms 2,367 ms

输入 3 3 3 1,467 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

30,854

5m: 0 / 1h: 30,854

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 21:36:14 ~ 21:36:24 (5.8s) 21:36:11 ~ 21:36:16 (4.2s) 21:37:28 ~ 21:37:38 (6.9s) 21:37:08 ~ 21:37:18 (5.9s) 21:38:05 ~ 21:38:15 (5.8s)

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) 3,865 ms 2,921 ms 5,067 ms 5,179 ms 2,605 ms

首文本(SSE delta) 3,866 ms 3,006 ms 5,446 ms 5,180 ms 2,606 ms

总耗时 4,621 ms 3,222 ms 6,091 ms 5,271 ms 2,606 ms

输入 3 3 3 3 3

输出 22 22 22 22 22

缓存创建

191

5m: 0 / 1h: 191

191

5m: 0 / 1h: 191

191

5m: 0 / 1h: 191

1,655

5m: 0 / 1h: 1,655

191

5m: 0 / 1h: 191

缓存读取 32,318 32,318 32,318 30,854 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 21:40:10 ~ 21:40:20 (6.2s) 21:37:23 ~ 21:37:28 (4.1s) 21:40:40 ~ 21:40:50 (5.1s) 21:39:54 ~ 21:40:04 (5.4s) 21:42:06 ~ 21:42:11 (3.0s)

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) 4,182 ms 2,881 ms 3,824 ms 4,540 ms 1,819 ms

首文本(SSE delta) 4,288 ms 2,889 ms 3,904 ms 4,541 ms 1,820 ms

总耗时 4,365 ms 3,095 ms 4,320 ms 4,620 ms 2,011 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

5m: 0 / 1h: 55

缓存读取 32,509 32,509 32,509 32,509 32,509

响应

2025-08

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 21:43:43 ~ 21:43:53 (7.0s) 21:38:59 ~ 21:39:04 (3.8s) 21:42:04 ~ 21:42:14 (7.9s) 21:43:27 ~ 21:43:37 (7.1s) 21:44:51 ~ 21:44:56 (4.6s)

提示词

查看本步 Prompt

请用中文介绍你自己是什么产品或助手，由谁开发，50 字以内。

HTTP 头(TTFB) 5,009 ms 1,870 ms 6,007 ms 6,314 ms 2,333 ms

首文本(SSE delta) 5,010 ms 1,871 ms 6,557 ms 6,315 ms 2,334 ms

总耗时 5,810 ms 2,850 ms 7,050 ms 6,411 ms 2,388 ms

输入 3 3 3 3 3

输出 54 55 26 55 49

缓存创建

5m: 0 / 1h: 43

5m: 0 / 1h: 35

5m: 0 / 1h: 43

缓存读取 32,566 32,564 32,564 32,566 32,564

输出速度 67.5 tok/s 56.2 tok/s - - -

响应

我是 Claude，由 Anthropic 开发的 AI 助手。当前版本为 Claude Sonnet 4.6，擅长编程、分析、写作等任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我基于大型语言模型构建，旨在提供安全、有益的对话和任务协助。

我是 Claude，由 Anthropic 公司开发的 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，旨在提供安全、有益的对话与任务协助。

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。

步骤 5: knowledge_recall text

时间 21:45:13 ~ 21:45:23 (5.7s) 21:40:05 ~ 21:40:10 (4.7s) 21:43:56 ~ 21:44:06 (5.3s) 21:44:58 ~ 21:45:03 (4.9s) 21:46:33 ~ 21:46:43 (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) 2,124 ms 1,349 ms 2,790 ms 4,152 ms 1,582 ms

首文本(SSE delta) 3,096 ms 1,350 ms 2,937 ms 4,153 ms 1,583 ms

总耗时 3,936 ms 3,689 ms 4,587 ms 4,200 ms 4,036 ms

输入 3 3 3 3 3

输出 205 204 204 205 205

缓存创建

8,052

5m: 0 / 1h: 8,052

8,043

5m: 0 / 1h: 8,043

8,014

5m: 0 / 1h: 8,014

8,045

5m: 0 / 1h: 8,045

8,045

5m: 0 / 1h: 8,045

缓存读取 0 0 0 0 0

输出速度 burst 87.2 tok/s 123.6 tok/s burst 83.6 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 21:48:13 ~ 21:48:28 (13.7s) 21:41:25 ~ 21:41:50 (21.0s) 21:45:31 ~ 21:45:56 (21.6s) 21:48:15 ~ 21:48:30 (12.6s) 21:50:25 ~ 21:50:45 (15.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=316960368287717399380724633934407612744342156693871233121843271755177305

HTTP 头(TTFB) 1,548 ms 1,746 ms 4,922 ms 6,629 ms 1,032 ms

首文本(SSE delta) 1,549 ms 1,747 ms 5,743 ms 6,630 ms 3,122 ms

总耗时 12,175 ms 19,995 ms 20,644 ms 11,901 ms 14,503 ms

输入 164 164 164 164 164

输出 978 1,703 1,282 1,063 1,124

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 92.0 tok/s 93.3 tok/s 86.0 tok/s burst 98.8 tok/s

响应

I'll carefully index every character in:

`316960368287717399380724633934407612744342156693871233121843271755177305`

**Length count (grouped in 10s):**
```
3169603682  → 10
8771739938  → 10
072463393

## Counting digits in s

**s** = `316960368287717399380724633934407612744342156693871233121843271755177305`

### Step 1 – Length (groups of 10)

```
3169603682  8771739938  0724633934  4076127443
4215

## Counting digits in s

**s** = `316960368287717399380724633934407612744342156693871233121843271755177305`

### Length
Grouping in 10s:
`3169603682` `8771739938` `0724633934` `4076127443` `4215669387

```json
{"n":72,"c":[4,11,5]}
```

**Verification walkthrough:**

```
s = 316960368287717399380724633934407612744342156693871233121843271755177305
     ^         ^         ^         ^         ^

```json
{"n":72,"c":[4,11,5]}
```

**Verification walkthrough:**

```
s = 316960368287717399380724633934407612744342156693871233121843271755177305
     ^    ^              ^^  ^      ^  ^   ^      ^ ^

总计

总输入(含缓存) 138,233 138,212 138,183 138,218 138,222

总输出 1,272 1,997 1,547 1,358 1,413

总缓存创建

40,661

5m: 0 / 1h: 40,661

40,642

5m: 0 / 1h: 40,642

40,613

5m: 0 / 1h: 40,613

40,646

5m: 0 / 1h: 40,646

40,652

5m: 0 / 1h: 40,652

总缓存读取 97,393 97,391 97,391 95,929 97,391

输出速度 90.3 tok/s 91.0 tok/s 89.8 tok/s burst 96.1 tok/s

总执行时间 44.7s 41.4s 53.2s 40.1s 38.3s

总墙钟时间

14m 5.1s

含步骤间等待

7m 19.0s

含步骤间等待

11m 27.0s

含步骤间等待

14m 5.0s

含步骤间等待

16m 18.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2013

Anthropic: Claude Sonnet 4.6

$0.2121

Anthropic: Claude Sonnet 4.6

$0.2053

Anthropic: Claude Sonnet 4.6

$0.2065

Anthropic: Claude Sonnet 4.6

$0.2034

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%) direct (100%) 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%) Anthropic 直连 (100%)

CDN cloudflare - - - cloudflare

ID 格式 msg_01XWv9Mt69XseKBiLSN9sLGT msg_015cKsrXdJ4LcEMcGxEd9RjW msg_012JkFrF84rtLfkHp1JDrTdE msg_01QKCGPYHmqjQDHV5YZMxd1z msg_01V2rWWvS9ZrLCvrQA6LqzZd

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	cf-ray: a124e880dc87f5df-AMS server: cloudflare x-oneapi-request-id: 202606271348153314225768268d9d6R07wb8xo	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	a124e880dc87f5df-AMS
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_011CcTrsLBpTLe9cQqFK25gz
cc专用	server: Nginx x-oneapi-request-id: 202606271345322223913038268d9d6iTjgyayT	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-k7ftlc0abs-101085
cc-Zeus	via: 1.1 Caddy cf-ray: a124ebb8d910b6a2-LAX server: cloudflare x-oneapi-request-id: 20260627215026713950623h5pnCJnw	id_format: msg_prefix via_chain: ["1.1 Caddy"] 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-rayhdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a124ebb8d910b6a2-LAX

协议指纹评分（机器自动）

总分 / 100 95 100 97 94 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 10.0 /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 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 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 8.2 /10 ×5 10.0 /10 ×5 8.8 /10 ×5 9.2 /10 ×5 10.0 /10 ×5

流式投递 4.2 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 10.0 /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 95 v3.25.0 基准（参考） 97 v3.25.0 94 v3.25.0 98 v3.25.0

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 5947 ms 5869 ms 7725 ms 5949 ms 4652 ms

输出速度 90.3 tok/s 91.0 tok/s 89.8 tok/s burst 96.1 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 ⓘ

$0.2013

$0.2121

$0.2053

$0.2065

$0.2034

缓存读取占比 ⓘ 70% 70% 70% 69% 70%

揭晓

通道

服务商 OmniaKey ↗

通道O-Max

官方基准

服务商 duckcoding ↗

通道cc专用

服务商 TopRouterCN ↗

通道O-Max

服务商 PackyCode ↗

通道cc-Zeus

返回列表