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

2026/06/22 运行日期

claude-sonnet-4-6 官方基准

2 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

TopRouterCN

O-Max

PackyCode

cc-Zeus

AnyAIGC

M-

luminai

M-Kiro/Antigravity

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 23:29:49 ~ 23:29:54 (4.8s) 23:29:55 ~ 23:30:00 (4.1s) 23:29:53 ~ 23:30:03 (6.7s) 23:29:51 ~ 23:30:01 (7.8s) 23:29:57 ~ 23:30:02 (3.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,923 ms 2,528 ms 3,332 ms 4,699 ms 2,267 ms

首文本(SSE delta) 3,925 ms 2,529 ms 3,334 ms 4,701 ms 2,269 ms

总耗时 3,995 ms 2,530 ms 3,887 ms 4,701 ms 2,415 ms

输入 1,467 3 3 1,443 3

输出 5 5 5 1 5

缓存创建

30,854

5m: 0 / 1h: 30,854

32,318

5m: 0 / 1h: 32,318

32,346

5m: 32,346 / 1h: 0

4,111

5m: 4,111 / 1h: 0

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 0 23,299 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 23:33:10 ~ 23:33:20 (9.7s) 23:33:23 ~ 23:33:33 (5.5s) 23:31:10 ~ 23:31:20 (8.1s) 23:33:25 ~ 23:33:35 (6.0s) 23:31:47 ~ 23:31:52 (3.2s)

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) 8,873 ms 2,085 ms 2,451 ms 4,482 ms 1,897 ms

首文本(SSE delta) 8,874 ms 2,086 ms 2,452 ms 4,483 ms 1,900 ms

总耗时 8,980 ms 3,241 ms 3,158 ms 4,483 ms 2,257 ms

输入 3 3 3 1,452 3

输出 22 22 22 24 22

缓存创建

1,655

5m: 0 / 1h: 1,655

191

5m: 0 / 1h: 191

191

5m: 191 / 1h: 0

4,137

5m: 4,137 / 1h: 0

191

5m: 0 / 1h: 191

缓存读取 30,854 32,318 32,346 23,445 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 23:35:54 ~ 23:36:14 (18.6s) 23:34:46 ~ 23:34:56 (5.0s) 23:33:57 ~ 23:34:07 (5.3s) 23:35:01 ~ 23:35:16 (10.4s) 23:34:52 ~ 23:34:57 (3.3s)

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) 17,635 ms 3,011 ms 2,845 ms 4,355 ms 2,154 ms

首文本(SSE delta) 17,637 ms 3,012 ms 2,846 ms 4,746 ms 2,156 ms

总耗时 17,727 ms 3,013 ms 4,050 ms 4,966 ms 2,318 ms

输入 3 3 3 1,455 3

输出 8 8 8 4 8

缓存创建

32,566

5m: 0 / 1h: 32,566

5m: 0 / 1h: 55

5m: 38 / 1h: 0

4,145

5m: 4,145 / 1h: 0

5m: 0 / 1h: 57

缓存读取 0 32,509 32,537 23,493 32,509

响应

2025-08

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 23:39:35 ~ 23:39:45 (9.0s) 23:38:52 ~ 23:39:02 (6.4s) 23:35:52 ~ 23:36:02 (5.9s) 23:36:31 ~ 23:36:41 (6.9s) 23:37:28 ~ 23:37:38 (5.0s)

提示词

查看本步 Prompt

用中文说说你的身份：什么产品？由谁打造？50 字以内。

HTTP 头(TTFB) 8,136 ms 3,957 ms 3,585 ms 5,510 ms 3,630 ms

首文本(SSE delta) 8,137 ms 3,958 ms 3,586 ms 5,650 ms 3,631 ms

总耗时 8,216 ms 4,109 ms 4,643 ms 5,910 ms 3,991 ms

输入 3 3 3 1,457 3

输出 46 65 42 56 24

缓存创建

1,751

5m: 0 / 1h: 1,751

5m: 0 / 1h: 43

5m: 35 / 1h: 0

4,152

5m: 4,152 / 1h: 0

5m: 0 / 1h: 35

缓存读取 30,854 32,564 32,575 23,530 32,566

输出速度 - - 39.7 tok/s - -

响应

我是 Claude，由 Anthropic 公司打造的 AI 助手。当前运行的模型版本为 claude-sonnet-4-6。

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、有帮助的回应。

我是 Claude，由 Anthropic 开发的 AI 助手。我致力于提供安全、有益的对话与协助。

我是 Claude，一个由 Anthropic 开发的 AI 驱动开发环境，基于 Claude Sonnet 4.6 模型，帮助开发者完成软件工程任务。

我是 Claude，由 Anthropic 开发的 AI 助手。

步骤 5: knowledge_recall text

时间 23:41:42 ~ 23:41:52 (9.7s) 23:42:57 ~ 23:43:12 (10.3s) 23:39:54 ~ 23:40:19 (20.0s) 23:39:45 ~ 23:39:55 (6.5s) 23:40:28 ~ 23:40:33 (4.8s)

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) 8,823 ms 4,459 ms 3,410 ms 3,240 ms 1,506 ms

首文本(SSE delta) 8,824 ms 4,460 ms 16,937 ms 5,369 ms 1,508 ms

总耗时 8,870 ms 6,834 ms 18,757 ms 5,443 ms 3,840 ms

输入 3 3 3 375 3

输出 204 205 1,032 210 204

缓存创建

8,040

5m: 0 / 1h: 8,040

8,061

5m: 0 / 1h: 8,061

8,041

5m: 8,041 / 1h: 0

1,068

5m: 1,068 / 1h: 0

8,014

5m: 0 / 1h: 8,014

缓存读取 0 0 0 6,054 0

输出速度 burst 86.4 tok/s burst burst 87.5 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 23:43:32 ~ 23:43:58 (23.1s) 23:46:50 ~ 23:47:05 (12.1s) 23:42:35 ~ 23:42:40 (3.7s) 23:42:19 ~ 23:42:29 (5.3s) 23:41:51 ~ 23:42:06 (14.1s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=753357155013928712503537800371695137984142002190765177682565107529513692

HTTP 头(TTFB) 11,879 ms 1,472 ms 1,811 ms 1,998 ms 1,416 ms

首文本(SSE delta) 11,880 ms 1,473 ms 1,829 ms 2,218 ms 1,418 ms

总耗时 22,395 ms 10,752 ms 2,430 ms 2,684 ms 13,006 ms

输入 164 164 166 188 164

输出 1,343 834 18 14 961

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 - 0

输出速度 127.7 tok/s 89.9 tok/s - - 82.9 tok/s

响应

**Counting the string:**

`753357155013928712503537800371695137984142002190765177682565107529513692`

**Length:** Splitting into groups of 10:
`7533571550` | `1392871250` | `3537800371` | `6951379841`

```json
{"n":72,"c":[8,10,6]}
```

**Verification walkthrough:**

| digit | positions (1-indexed) | count |
|-------|-----------------------|-------|
| **0** | 10,20,26,27,43,44,48,62 | **8** |
| **7*

```json
{"n":72,"c":[5,6,7]}
```

{"n":72,"c":[4,6,5]}

```json
{"n":72,"c":[8,10,6]}
```

**Verification walkthrough:**

String (indexed):
```
7 5 3 3 5 7 1 5 5 0  ← pos 1-10
1 3 9 2 8 7 1 2 5 0  ← pos 11-20
3 5 3 7 8 0 0 3 7 1  ← pos 21-30
6 9 5 1 3 7 9

总计

总输入(含缓存) 138,217 138,238 138,290 123,804 138,187

总输出 1,628 1,139 1,127 309 1,224

总缓存创建

74,866

5m: 0 / 1h: 74,866

40,668

5m: 0 / 1h: 40,668

40,651

5m: 40,651 / 1h: 0

17,613

5m: 17,613 / 1h: 0

40,615

5m: 0 / 1h: 40,615

总缓存读取 61,708 97,391 97,458 99,821 97,393

输出速度 127.7 tok/s 89.2 tok/s 39.7 tok/s burst 83.7 tok/s

总执行时间 1m 14.9s 43.3s 49.8s 43.0s 34.2s

总墙钟时间

14m 8.2s

含步骤间等待

17m 10.0s

含步骤间等待

12m 47.0s

含步骤间等待

12m 38.1s

含步骤间等待

12m 9.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3286

Anthropic: Claude Sonnet 4.6

$0.1993

Anthropic: Claude Sonnet 4.6

$0.1420

Anthropic: Claude Sonnet 4.6

$0.0950

Anthropic: Claude Sonnet 4.6

$0.2004

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

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

CDN - cloudflare - cloudflare -

ID 格式 msg_01Mk7TnGu61Zwq2eMJEpFn9y msg_01GEeZiKkt5o9GQsXkfmjG4n 自定义 msg_01tAYhufTMdoqnDuefbcAhDVma msg_014QFJQA4XYovE1SPBRQHALi

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/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-tlfecbjbfy-057795
cc-Zeus	via: 1.1 Caddy cf-ray: a0fc636429602ae9-LAX server: cloudflare x-oneapi-request-id: 20260622234652208637807goqJK203	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	a0fc636429602ae9-LAX
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	—
M-Kiro/Antigravity	cf-ray: a0fc5ccf3d6a3466-AMS server: cloudflare x-oneapi-request-id: 202606221542227007755758268d9d6pxqddJk5	id_format: msg_prefix extra_fields: ["container"] has_service_tier: true has_inference_geo: false 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.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0fc5ccf3d6a3466-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_011CcJZ1U1TWkquy7Vyh2bRt

协议指纹评分（机器自动）

总分 / 100 83 98 85 74 99

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 4.5 /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 0.0 /10 ×15 0.0 /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 7.7 /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 0.0 /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 0.0 /10 ×5 10.0 /10 ×5

延迟基线 0.5 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 9.3 /10 ×5 10.0 /10 ×5

流式投递 0.0 /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 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 83 v3.24.0 98 v3.24.0 85 v3.24.0 74 v3.24.0 基准（参考）

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 11697 ms 5080 ms 6154 ms 4698 ms 4638 ms

输出速度 127.7 tok/s 89.2 tok/s 39.7 tok/s burst 83.7 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 ⓘ

$0.3286

$0.1993

$0.1420

$0.0950

$0.2004

缓存读取占比 ⓘ 45% 70% 70% 81% 70%

揭晓

通道

服务商 TopRouterCN ↗

通道O-Max

服务商 PackyCode ↗

通道cc-Zeus

服务商 AnyAIGC ↗

通道M-

服务商 luminai ↗

通道M-Kiro/Antigravity

官方基准

返回列表