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。

4 通道 · 6 步骤

2026/06/23 运行日期

claude-sonnet-4-6 官方基准

2 任务失败执行不完整

已揭晓揭晓进度

⚠ 1 通道不可测 · 已从主对比表移除点击展开原因

DBAI / M-
```
评测失败
```

步骤 / 指标

SSSAiCode

O-Max-L2

官方基准

ClaudeZZ

R-kiro

DawAPI

基本信息

服务名 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

客户端 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 17:02:34 ~ 17:02:39 (3.5s) 17:02:38 ~ 17:02:48 (5.2s) 17:02:30 ~ 17:02:50 (17.2s) 17:02:36 ~ 17:02:46 (6.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,030 ms 3,672 ms 15,974 ms 4,767 ms

首文本(SSE delta) 2,102 ms 3,675 ms 15,976 ms 4,769 ms

总耗时 2,168 ms 3,922 ms 16,021 ms 4,879 ms

输入 3 3 3 3

输出 5 5 5 5

缓存创建

31,987

5m: 0 / 1h: 31,987

32,318

5m: 0 / 1h: 32,318

21,984

5m: 0 / 1h: 21,984

32,346

5m: 32,346 / 1h: 0

缓存读取 0 0 0 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 17:04:21 ~ 17:04:31 (9.5s) 17:04:41 ~ 17:04:46 (3.7s) 17:05:22 ~ 17:05:37 (12.0s) 17:04:49 ~ 17:04:59 (5.3s)

提示词

查看本步 Prompt

输出 3 行 metadata 描述你自己（当前 API 模型）。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。禁止任何额外字符：
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) 8,289 ms 2,260 ms 10,621 ms 4,087 ms

首文本(SSE delta) 8,539 ms 2,262 ms 10,622 ms 4,089 ms

总耗时 8,790 ms 2,663 ms 11,225 ms 4,486 ms

输入 3 3 3 3

输出 22 22 22 22

缓存创建

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

196

5m: 0 / 1h: 196

189

5m: 189 / 1h: 0

缓存读取 31,987 32,318 21,984 32,346

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

步骤 3: cutoff text

时间 17:08:07 ~ 17:08:13 (4.0s) 17:07:17 ~ 17:07:22 (4.5s) 17:08:10 ~ 17:08:20 (7.8s) 17:08:51 ~ 17:09:01 (9.5s)

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) 2,967 ms 3,164 ms 6,359 ms 4,563 ms

首文本(SSE delta) 3,076 ms 3,381 ms 6,360 ms 4,838 ms

总耗时 3,183 ms 3,492 ms 6,585 ms 5,612 ms

输入 3 3 3 3

输出 8 8 8 8

缓存创建

5m: 0 / 1h: 55

5m: 0 / 1h: 62

5m: 38 / 1h: 0

缓存读取 32,176 32,507 22,180 32,535

响应

2025-04

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 17:10:19 ~ 17:10:29 (5.7s) 17:10:48 ~ 17:10:58 (5.0s) 17:11:03 ~ 17:11:13 (9.0s) 17:11:42 ~ 17:11:52 (8.7s)

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) 4,260 ms 2,509 ms 5,543 ms 5,010 ms

首文本(SSE delta) 4,907 ms 2,511 ms 5,545 ms 5,414 ms

总耗时 4,988 ms 4,084 ms 6,867 ms 6,601 ms

输入 3 3 3 3

输出 52 68 63 46

缓存创建

5m: 0 / 1h: 35

5m: 0 / 1h: 43

5m: 0 / 1h: 53

5m: 41 / 1h: 0

缓存读取 32,231 32,562 22,242 32,573

输出速度 - 43.2 tok/s 47.7 tok/s 38.8 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我的设计目标是安全、有益且诚实地协助用户完成各类任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助你完成编程、写作、分析、问答等各类任务，致力于提供安全、准确、有帮助的回应。

我是 Claude，由 Anthropic 开发的 AI 助手。我是 Claude Code，专为软件工程任务设计，可帮助你编写代码、调试问题、分析代码库等。

我叫 Claude，是由 Anthropic 开发的 AI 助手，当前运行的模型版本为 claude-sonnet-4-6。

步骤 5: knowledge_recall text

时间 17:14:07 ~ 17:14:12 (4.6s) 17:14:57 ~ 17:15:02 (4.4s) 17:13:23 ~ 17:13:33 (6.7s) 17:13:07 ~ 17:13:17 (5.9s)

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) 1,520 ms 1,539 ms 3,112 ms 3,234 ms

首文本(SSE delta) 3,739 ms 1,540 ms 3,114 ms 3,235 ms

总耗时 3,820 ms 3,379 ms 5,677 ms 5,077 ms

输入 3 3 3 3

输出 198 204 204 204

缓存创建

7,707

5m: 0 / 1h: 7,707

8,062

5m: 0 / 1h: 8,062

6,114

5m: 0 / 1h: 6,114

8,049

5m: 8,049 / 1h: 0

缓存读取 0 0 2,136 0

输出速度 - 110.9 tok/s 79.6 tok/s 110.7 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

步骤 6: digit_count text

时间 17:17:18 ~ 17:17:23 (1.3s) 17:17:08 ~ 17:17:28 (15.0s) 17:14:35 ~ 17:14:55 (16.7s) 17:17:14 ~ 17:17:44 (26.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=479703204695175476767920993079834132124855861640086246223931846767696014

HTTP 头(TTFB) - 1,712 ms 3,357 ms 15,611 ms

首文本(SSE delta) - 1,714 ms 3,359 ms 24,120 ms

总耗时 - 13,851 ms 15,773 ms 25,894 ms

输入 - 164 3 193

输出 - 970 1,554 1,075

缓存创建

5m: 0 / 1h: 0

3,821

5m: 0 / 1h: 3,821

5m: 0 / 1h: 0

缓存读取 - 0 2,136 0

输出速度 - 79.9 tok/s 125.2 tok/s burst

响应

失败

评测失败

```json
{"n":72,"c":[7,9,8]}
```

**Verification breakdown** (groups of 10):

| Group | Digits | 0s | 7s | 9s |
|-------|--------|----|----|----|
| 4797032046 | 10 | 2 | 2 | 1 |
| 9517547676 | 10 | 0

I'll count by indexing each character in s = "479703204695175476767920993079834132124855861640086246223931846767696014"

**Indexing (1-based):**
```
 1: 4   2: 7   3: 9   4: 7   5: 0   6: 3   7: 2   8

```python
s = "479703204695175476767920993079834132124855861640086246223931846767696014"
n = len(s)
c0 = s.count('0')
c7 = s.count('7')
c9 = s.count('9')
```

**Results:**

- `n = 72`
- `0` appears **

总计

总输入(含缓存)

136,382

仅前 6 步

138,233

102,926

138,325

总输出

285

仅前 6 步

1,277

1,856

1,360

总缓存创建

39,973

5m: 0 / 1h: 39,973

仅前 6 步

40,667

5m: 0 / 1h: 40,667

32,230

5m: 0 / 1h: 32,230

40,663

5m: 40,663 / 1h: 0

总缓存读取

96,394

仅前 6 步

97,387

70,678

97,454

输出速度

仅前 6 步

79.9 tok/s

111.7 tok/s

82.5 tok/s

总执行时间

28.8s

仅前 6 步

37.8s

1m 9.5s

1m 2.3s

总墙钟时间

14m 49.0s

含步骤间等待

仅前 6 步

14m 50.0s

含步骤间等待

12m 25.1s

含步骤间等待

15m 8.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1831

Anthropic: Claude Sonnet 4.6 (部分)

$0.2014

Anthropic: Claude Sonnet 4.6

$0.1700

Anthropic: Claude Sonnet 4.6

$0.1456

Anthropic: Claude Sonnet 4.6

通道指纹

来源步骤 knowledge_recall digit_count digit_count digit_count

平台 direct (100%) direct (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 - - - -

ID 格式 msg_01RXH2ojFYuQEEWZL82VUpBq msg_014HwZfrBYXh2vx1rscoXEG7 msg_01UAwK7RUen6yU7PAQDxyGGq msg_01DWnEukKE1VMRCshTFF2gKF

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-L2	cf-ray: a102617a68c65027-LAX server: Photon-Edge set-cookie: _cfuvid=4Fky1bBi5CtM52WJRzh0xv1eShPTOYjgilYpZbARYE8-1782206048.3908663-1.0.1.1-ADpJabyzsMAkHAA_GehcopHCLn1S1b3.hHguNAyTblk; 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	1782206048343-kt4cyukxd8bha req_011CcKwFSjWxK38A5pgYUEak a102617a68c65027-LAX
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_011CcKwUpFwKjzTiDuo3XMJX
R-kiro	via: 1.1 google cf-ray: df0c21d13ccb21958eeedbe935e928bd-IAD server: nginx	id_format: msg_prefix via_chain: ["1.1 google"] extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_df0c21d13ccb21958eeedbe935e928bd df0c21d13ccb21958eeedbe935e928bd-IAD
cc	server: cdn x-new-api-version: v1.0.0-rc.14 x-oneapi-request-id: 202606230917156441929398268d9d6qPXWJDmE	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	—

协议指纹评分（机器自动）

总分 / 100 100 90 88

基线状态已对照基线已对照基线已对照基线

模型匹配 — 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

sliding 5m — 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

缓存 TTL 一致 — 10.0 /10 ×15 10.0 /10 ×15 0.0 /10 ×15

缓存命中比 — 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

身份(结构) — 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

身份(自由) — 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7

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

Req-ID 透传 — 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

service_tier — 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

SDK 一致 — 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

系统提示纯净 — 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8

延迟基线 — 10.0 /10 ×5 0.0 /10 ×5 7.5 /10 ×5

流式投递 — 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

综合结论

与基线相似度执行失败，不评分基准（参考） 90 v3.24.0 88 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 4590 ms 5232 ms 10358 ms 8758 ms

输出速度 - 79.9 tok/s 111.7 tok/s 82.5 tok/s

完成步骤 5/6 成功 · 1 失败 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 ⓘ

$0.1831 (部分)

$0.2014

$0.1700

$0.1456

缓存读取占比 ⓘ 71% (部分) 70% 69% 70%

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

官方基准

服务商 ClaudeZZ ↗

通道R-kiro

服务商 DawAPI ↗

通道cc

返回列表