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

2026/06/26 运行日期

claude-haiku-4-5-20251001 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

SSSAiCode

O-Max-L2

SAIAi

O-Max

AnyAIGC

M-

官方基准

ClaudeCN

O-Max

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001

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 18:59:04 ~ 18:59:09 (3.8s) 18:59:08 ~ 18:59:38 (25.2s) 18:59:04 ~ 18:59:09 (4.4s) 18:59:10 ~ 18:59:50 (36.0s) 18:59:06 ~ 18:59:16 (5.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,106 ms 20,788 ms 2,974 ms 27,360 ms 3,267 ms

首文本(SSE delta) 2,409 ms 20,789 ms 2,976 ms 27,362 ms 3,269 ms

总耗时 2,565 ms 21,153 ms 2,976 ms 27,779 ms 3,886 ms

输入 10 10 8,600 10 10

输出 46 54 2 55 50

缓存创建

32,790

5m: 0 / 1h: 32,790

33,122

5m: 0 / 1h: 33,122

5m: 0 / 1h: 0

33,122

5m: 0 / 1h: 33,122

缓存读取 0 0 0 33,122 0

输出速度 - 148.4 tok/s - 131.9 tok/s 81.0 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 19:02:52 ~ 19:03:02 (7.3s) 19:02:19 ~ 19:02:29 (7.4s) 19:02:15 ~ 19:02:25 (6.3s) 19:01:28 ~ 19:01:38 (5.2s) 19:03:16 ~ 19:03:26 (6.3s)

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。自报身份，输出严格遵守以下 3 行格式（不允许 markdown、代码块、额外文本）：
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) 3,719 ms 4,667 ms 4,125 ms 2,515 ms 2,909 ms

首文本(SSE delta) 6,160 ms 4,839 ms 4,126 ms 2,517 ms 2,910 ms

总耗时 6,414 ms 6,447 ms 4,133 ms 4,168 ms 5,161 ms

输入 10 10 8,799 10 10

输出 292 189 25 185 214

缓存创建

240

5m: 0 / 1h: 240

248

5m: 0 / 1h: 248

249

5m: 0 / 1h: 249

244

5m: 0 / 1h: 244

缓存读取 32,790 33,122 0 33,122 33,122

输出速度 burst 117.5 tok/s - 112.1 tok/s 95.1 tok/s

响应

vendor: anthropic
brand: claude
model: claude-3-5-sonnet-20241022

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic  
brand: claude  
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

步骤 3: cutoff text

时间 19:05:05 ~ 19:05:10 (4.5s) 19:04:16 ~ 19:04:22 (3.6s) 19:03:49 ~ 19:03:59 (5.3s) 19:03:26 ~ 19:03:31 (3.2s) 19:06:02 ~ 19:06:13 (8.7s)

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) 2,703 ms 2,131 ms 2,953 ms 1,744 ms 4,190 ms

首文本(SSE delta) 3,090 ms 2,132 ms 2,954 ms 1,752 ms 4,192 ms

总耗时 3,517 ms 2,570 ms 3,837 ms 2,171 ms 7,415 ms

输入 10 10 8,856 10 10

输出 65 66 5 68 66

缓存创建

332

5m: 0 / 1h: 332

231

5m: 0 / 1h: 231

227

5m: 0 / 1h: 227

33,620

5m: 0 / 1h: 33,620

缓存读取 33,030 33,370 0 33,371 0

输出速度 152.2 tok/s 150.7 tok/s - 162.3 tok/s 20.5 tok/s

响应

2024-04

2025-02

2025-05

2025-02

2025-02

步骤 4: identity_free text

时间 19:08:11 ~ 19:08:21 (5.5s) 19:07:56 ~ 19:08:06 (7.7s) 19:05:51 ~ 19:05:56 (4.1s) 19:07:07 ~ 19:07:17 (6.4s) 19:09:42 ~ 19:09:52 (9.6s)

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) 3,076 ms 2,929 ms 2,888 ms 3,218 ms 3,430 ms

首文本(SSE delta) 4,353 ms 2,931 ms 2,889 ms 3,219 ms 3,431 ms

总耗时 4,714 ms 6,668 ms 2,889 ms 5,426 ms 7,313 ms

输入 10 10 8,893 10 10

输出 219 507 22 324 347

缓存创建

107

5m: 0 / 1h: 107

100

5m: 0 / 1h: 100

108

5m: 0 / 1h: 108

108

5m: 0 / 1h: 108

缓存读取 33,362 33,601 0 33,598 33,620

输出速度 - 135.7 tok/s - 146.8 tok/s 89.4 tok/s

响应

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行编程、写作、分析、研究等各种任务，并能使用多种工具完成复杂工作。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。可帮助编程、代码分析和问题解答。

我叫Claude，属于Anthropic开发的AI助手，旨在提供专业、安全的支持。

我是Claude，由Anthropic公司开发的AI助手。目前运行的是Claude Haiku 4.5模型。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助编写代码、分析问题、进行研究等多种任务。

步骤 5: knowledge_recall text

时间 19:11:23 ~ 19:11:43 (15.2s) 19:11:03 ~ 19:11:23 (19.1s) 19:08:31 ~ 19:08:41 (7.6s) 19:10:40 ~ 19:11:00 (18.3s) 19:12:03 ~ 19:12:28 (20.6s)

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) 1,551 ms 1,682 ms 3,092 ms 1,843 ms 2,529 ms

首文本(SSE delta) 13,958 ms 1,683 ms 3,093 ms 1,845 ms 2,530 ms

总耗时 14,315 ms 18,149 ms 4,128 ms 17,372 ms 19,362 ms

输入 10 10 9,513 10 10

输出 2,096 2,550 154 2,345 2,290

缓存创建

8,306

5m: 0 / 1h: 8,306

3,105

5m: 0 / 1h: 3,105

8,648

5m: 0 / 1h: 8,648

8,693

5m: 0 / 1h: 8,693

缓存读取 0 5,721 0 0 0

输出速度 burst 154.9 tok/s 148.8 tok/s 151.0 tok/s 136.1 tok/s

响应

```json
{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_pe

```json
{"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

{"facts":{"super_bowl_lviii":"UNKNOWN","oscars2024_best_picture":"UNKNOWN","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_peace_2024":"UNKNOWN","wo

```json
{"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

```json
{"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

步骤 6: digit_count text

时间 19:12:58 ~ 19:13:23 (22.0s) 19:13:34 ~ 19:13:49 (14.8s) 19:10:06 ~ 19:10:11 (3.2s) 19:14:52 ~ 19:15:12 (18.9s) 19:14:24 ~ 19:14:44 (16.0s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=385585355303655709462511645136733572671063220207120586765569724262228993

HTTP 头(TTFB) 1,389 ms 1,106 ms 1,667 ms 1,153 ms 1,682 ms

首文本(SSE delta) 1,390 ms 1,107 ms 1,668 ms 1,157 ms 1,684 ms

总耗时 21,261 ms 13,906 ms 1,668 ms 18,021 ms 14,843 ms

输入 139 193 166 193 193

输出 4,095 2,337 14 3,303 2,390

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 206.1 tok/s 182.6 tok/s - 195.9 tok/s 181.6 tok/s

响应

```json
{"n":72,"c":[6,7,4]}
```

```json
{"n":72,"c":[6,7,4]}
```

{"n":78,"c":[4,4,2]}

```json
{"n": 72, "c": [6, 7, 4]}
```

```json
{"n":72,"c":[6,7,4]}
```

总计

总输入(含缓存) 141,146 142,863 44,827 142,688 142,772

总输出 6,813 5,703 222 6,280 5,357

总缓存创建

41,775

5m: 0 / 1h: 41,775

36,806

5m: 0 / 1h: 36,806

9,232

5m: 0 / 1h: 9,232

75,787

5m: 0 / 1h: 75,787

总缓存读取 99,182 105,814 0 133,213 66,742

输出速度 204.9 tok/s 161.0 tok/s 148.8 tok/s 169.3 tok/s 134.0 tok/s

总执行时间 58.2s 1m 17.8s 30.9s 1m 28.0s 1m 6.3s

总墙钟时间

14m 19.4s

含步骤间等待

14m 41.3s

含步骤间等待

11m 7.4s

含步骤间等待

16m 2.1s

含步骤间等待

15m 38.3s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0964

Anthropic: Claude Haiku 4.5

$0.0853

Anthropic: Claude Haiku 4.5

$0.0459

Anthropic: Claude Haiku 4.5

$0.0565

Anthropic: Claude Haiku 4.5

$0.1284

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count digit_count

平台 direct (100%) direct (100%) direct (50%) direct (100%) one-api (100%)

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

CDN - - - - cloudflare

ID 格式 msg_01M7Rg3MnTYimLs5hoAcWs7K msg_01DPYHkdRZyfb5maFzcEMysX 自定义 msg_013WutSpdtvYCiD17qhVHyRA msg_01TDk5Hv1sLgVJtFXCJkXHSQ

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: a11bc7b76c102f71-LAX server: Photon-Edge set-cookie: _cfuvid=cIA_1QmvvgK_UU9NTMHM75PVmM962aaydgpc6ti_0LE-1782472380.065016-1.0.1.1-Kd43pW6_OO3NA05_CFhgbzEvUwAo5qfzdUzI0p4ub8Q; 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: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	1782472379444-zwysheetft5eo req_011CcRmjtwkCCyXUi6Vs2qvo a11bc7b76c102f71-LAX
O-Max	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: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcRmnWv1TfmR983kRFUjC
M-	server: nginx	id_format: non_standard has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	body:id:non_standardbody:model:date_suffixbody:usage.cache_fields	—
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: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcRmtEwVUy1UiHL4vr4DS
O-Max	cf-ray: a11bc9cf684d65f5-AMS server: cloudflare x-oneapi-request-id: 20260626191426391699569DMH7BUYh	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a11bc9cf684d65f5-AMS

协议指纹评分（机器自动）

总分 / 100 76 99 34 100 85

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 7.4 /10 ×20 8.7 /10 ×20 0.0 /10 ×20 10.0 /10 ×20 0.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 0.0 /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 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 10.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 7.7 /10 ×13

原生 msg-ID 10.0 /10 ×8 10.0 /10 ×8 0.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

系统提示纯净 5.0 /10 ×8 10.0 /10 ×8 0.0 /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 10.0 /10 ×7

知识截止 0.0 /10 ×7 10.0 /10 ×7 5.0 /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 0.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

inference_geo 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

延迟基线 9.9 /10 ×5 10.0 /10 ×5 9.8 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 10.0 /10 ×4 10.0 /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 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

流式投递未计入未计入未计入未计入未计入

综合结论

与基线相似度 76 v3.25.0 99 v3.25.0 34 v3.25.0 基准（参考） 85 v3.25.0

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 Thinking 存在 thinking_present
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 Thinking 体量 thinking_volume_match
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 自报身份一致 self_identity_consistency
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 8798 ms 11482 ms 3272 ms 12490 ms 9663 ms

输出速度 204.9 tok/s 161.0 tok/s 148.8 tok/s 169.3 tok/s 134.0 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 ⓘ

$0.0964

$0.0853

$0.0459

$0.0565

$0.1284

缓存读取占比 ⓘ 70% 74% 0% 93% 47%

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 SAIAi ↗

通道O-Max

服务商 AnyAIGC ↗

通道M-

官方基准

服务商 ClaudeCN ↗

通道O-Max

返回列表