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

2026/07/01 运行日期

claude-haiku-4-5-20251001 官方基准

1 通道分数 < 70 另含 1 任务失败

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

SSSAiCode / R-Kiro
```
评测失败
```

步骤 / 指标

SSSAiCode

O-Max-L2

官方基准

AIMZ

M-Max

SAIAi

O-Max

基本信息

服务名 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

客户端 claude-cli v2.1.195 claude-cli v2.1.195 claude-cli v2.1.195 claude-cli v2.1.195

偏差标红基于官方基准

步骤 1: ping text

时间 07:50:10 ~ 07:50:15 (3.2s) 07:50:12 ~ 07:50:17 (3.1s) 07:50:08 ~ 07:50:13 (4.5s) 07:50:06 ~ 07:50:11 (3.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,800 ms 1,675 ms 2,617 ms 1,377 ms

首文本(SSE delta) 2,141 ms 1,676 ms 3,048 ms 1,391 ms

总耗时 2,324 ms 2,111 ms 3,102 ms 1,911 ms

输入 10 10 10 10

输出 49 43 53 56

缓存创建

5,761

5m: 0 / 1h: 5,761

5m: 0 / 1h: 0

21,177

5m: 0 / 1h: 21,177

6,166

5m: 0 / 1h: 6,166

缓存读取 15,418 21,545 0 15,379

输出速度 - 98.9 tok/s - 107.7 tok/s

响应

pong

pong

pong

pong

步骤 2: identity text

时间 07:53:34 ~ 07:53:44 (7.9s) 07:52:08 ~ 07:52:13 (4.7s) 07:52:12 ~ 07:52:33 (15.1s) 07:53:27 ~ 07:53:32 (3.9s)

提示词

查看本步 Prompt

以下是对你自身的 metadata 探针。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。按以下 3 行回答自己的身份，每行一个 label，不要寒暄不要解释：
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) 1,746 ms 1,594 ms 1,836 ms 1,468 ms

首文本(SSE delta) 6,814 ms 1,596 ms 13,882 ms 1,470 ms

总耗时 7,191 ms 3,725 ms 13,967 ms 2,979 ms

输入 10 10 10 10

输出 536 211 1,350 144

缓存创建

253

5m: 0 / 1h: 253

247

5m: 0 / 1h: 247

257

5m: 0 / 1h: 257

260

5m: 0 / 1h: 260

缓存读取 21,179 21,545 21,177 21,545

输出速度 burst 99.1 tok/s burst 95.4 tok/s

响应

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

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

vendor: anthropic
brand: claude
model: I cannot determine the exact API model identifier from my system context.

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

步骤 3: cutoff text

时间 07:55:01 ~ 07:55:06 (3.1s) 07:53:21 ~ 07:53:26 (3.3s) 07:54:15 ~ 07:54:25 (7.9s) 07:55:50 ~ 07:55:55 (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) 1,670 ms 1,713 ms 2,169 ms 1,376 ms

首文本(SSE delta) 2,222 ms 1,715 ms 6,839 ms 1,377 ms

总耗时 2,396 ms 2,303 ms 6,887 ms 1,994 ms

输入 10 10 10 10

输出 85 62 532 66

缓存创建

578

5m: 0 / 1h: 578

234

5m: 0 / 1h: 234

1,390

5m: 0 / 1h: 1,390

184

5m: 0 / 1h: 184

缓存读取 21,432 21,792 21,434 21,805

输出速度 - 105.4 tok/s burst 107.0 tok/s

响应

2024-04

2025-02

2024-10

2025-02

步骤 4: identity_free text

时间 07:58:52 ~ 07:59:07 (10.2s) 07:55:35 ~ 07:55:50 (10.8s) 07:55:46 ~ 07:56:06 (18.4s) 07:58:24 ~ 07:58:34 (9.2s)

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) 1,633 ms 2,377 ms 2,002 ms 1,568 ms

首文本(SSE delta) 9,223 ms 2,379 ms 17,153 ms 1,570 ms

总耗时 9,401 ms 9,821 ms 17,207 ms 8,216 ms

输入 10 10 10 10

输出 1,266 740 2,727 635

缓存创建

127

5m: 0 / 1h: 127

100

5m: 0 / 1h: 100

574

5m: 0 / 1h: 574

108

5m: 0 / 1h: 108

缓存读取 22,010 22,026 22,824 21,989

输出速度 burst 99.4 tok/s burst 95.5 tok/s

响应

我是Claude，由Anthropic开发的AI助手。可以帮助编程、写作、分析等任务，努力提供准确有用的回应。

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手，具备代码协助、分析等能力，知识库截至2025年2月。

我是Claude，由Anthropic开发的AI助手。协助编程、写作、分析等多类任务，提供安全诚实的帮助。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行代码编程、文本创作、问题解答等各种任务。我致力于提供安全、有用和诚实的帮助。

步骤 5: knowledge_recall text

时间 08:00:43 ~ 08:00:58 (11.0s) 07:57:22 ~ 07:57:57 (33.3s) 07:58:46 ~ 07:59:06 (16.4s) 08:00:15 ~ 08:00:40 (22.5s)

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) 2,188 ms 1,493 ms 1,805 ms 1,307 ms

首文本(SSE delta) 9,861 ms 1,495 ms 15,115 ms 1,309 ms

总耗时 10,226 ms 32,303 ms 15,327 ms 21,464 ms

输入 10 10 10 10

输出 1,704 3,558 2,024 2,402

缓存创建

10,354

5m: 0 / 1h: 10,354

9,817

5m: 0 / 1h: 9,817

13,076

5m: 0 / 1h: 13,076

9,683

5m: 0 / 1h: 9,683

缓存读取 0 0 0 0

输出速度 burst 115.5 tok/s burst 119.2 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

```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 08:03:08 ~ 08:03:23 (14.4s) 08:01:26 ~ 08:01:56 (27.7s) 08:01:31 ~ 08:01:46 (12.9s) 08:01:49 ~ 08:02:09 (19.0s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=895801205958439515476022953520689468480311632521675882513902030949536846

HTTP 头(TTFB) 1,454 ms 939 ms 1,106 ms 1,811 ms

首文本(SSE delta) 1,503 ms 943 ms 1,107 ms 1,813 ms

总耗时 13,703 ms 26,785 ms 11,831 ms 18,040 ms

输入 139 193 139 193

输出 2,191 3,359 1,987 1,661

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 179.6 tok/s 130.0 tok/s 185.3 tok/s 102.4 tok/s

响应

```json
{"n":72,"c":[8,2,8]}
```

```json
{"n":72,"c":[8,2,8]}
```

```json
{"n": 72, "c": [8, 2, 8]}
```

```json
{"n": 74, "c": [8, 2, 9]}
```

总计

总输入(含缓存) 97,301 97,549 102,098 97,362

总输出 5,831 7,973 8,673 4,964

总缓存创建

17,073

5m: 0 / 1h: 17,073

10,398

5m: 0 / 1h: 10,398

36,474

5m: 0 / 1h: 36,474

16,401

5m: 0 / 1h: 16,401

总缓存读取 80,039 86,908 65,435 80,718

输出速度 179.6 tok/s 118.6 tok/s 185.3 tok/s 108.7 tok/s

总执行时间 49.8s 1m 22.9s 1m 15.3s 1m 0.6s

总墙钟时间

13m 13.0s

含步骤间等待

11m 44.1s

含步骤间等待

11m 38.1s

含步骤间等待

12m 3.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0587

Anthropic: Claude Haiku 4.5

$0.0618

Anthropic: Claude Haiku 4.5

$0.0957

Anthropic: Claude Haiku 4.5

$0.0536

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 direct (100%) direct (100%) 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 -

ID 格式 msg_01BasgfDG5GZpzKUQiKQyzHp msg_019fZeSMQ7Cx3rUmsiQNeQ3K msg_01QWhb9PLm7XxbLiipTJswRQ msg_01KHqDKRxjzzXXt8hbA8PPZQ

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: a14125621e2150af-LAX server: Photon-Edge	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	1782864189620-r83an6e4fkrg4 req_011CcaMhsoDcHCvDHuFp42XE a14125621e2150af-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: 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_011CcaMaN4rdZYAS2CiWXTm3
M-Max	cf-ray: a141230509270eb0-AMS server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202607010001330149042608268d9d6p7QCRp7k	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-new-api-versionhdr: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	a141230509270eb0-AMS
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_011CcaMc5ocqCL49BYV72DYY

协议指纹评分（机器自动）

总分 / 100 85 99 84 100

基线状态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 7.3 /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

延迟基线 10.0 /10 ×15 10.0 /10 ×15 7.1 /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

缓存连续性 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

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

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

身份(结构) 6.7 /10 ×7 10.0 /10 ×7 6.7 /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

service_tier 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

知识截止 0.0 /10 ×5 10.0 /10 ×5 5.0 /10 ×5 10.0 /10 ×5

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

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

Req-ID 透传 10.0 /10 ×2 10.0 /10 ×2 0.0 /10 ×2 10.0 /10 ×2

SDK 一致 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

综合结论

与基线相似度 85 v3.26.0 基准（参考） 84 v3.26.0 100 v3.26.0

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 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 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 7540 ms 12841 ms 11387 ms 9101 ms

输出速度 179.6 tok/s 118.6 tok/s 185.3 tok/s 108.7 tok/s

完成步骤 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 ⓘ

$0.0587

$0.0618

$0.0957

$0.0536

缓存读取占比 ⓘ 82% 89% 64% 83%

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

官方基准

服务商 AIMZ ↗

通道M-Max

服务商 SAIAi ↗

通道O-Max

返回列表