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

2026/06/18 运行日期

claude-haiku-4-5-20251001 官方基准

3 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

LinkAPI

O-Max

FoxCode

R-ultra

HongMaCC

u-cc

官方基准

SSSAiCode

O-Max-L2

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 07:12:19 ~ 07:12:29 (5.0s) 07:12:17 ~ 07:12:22 (4.7s) 07:12:21 ~ 07:12:26 (4.0s) 07:12:25 ~ 07:12:30 (3.6s) 07:12:23 ~ 07:12:28 (3.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,016 ms 2,551 ms 2,745 ms 2,165 ms 2,152 ms

首文本(SSE delta) 3,020 ms 2,552 ms 2,747 ms 2,167 ms 2,488 ms

总耗时 3,741 ms 3,219 ms 2,747 ms 2,450 ms 2,523 ms

输入 10 33,019 0 10 10

输出 57 44 19 41 43

缓存创建

33,122

5m: 0 / 1h: 33,122

5m: 0 / 1h: 0

33,122

5m: 0 / 1h: 33,122

32,860

5m: 0 / 1h: 32,860

缓存读取 0 0 0 0 0

输出速度 79.1 tok/s 66.0 tok/s - 144.9 tok/s burst

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 07:14:40 ~ 07:14:50 (6.4s) 07:15:14 ~ 07:15:19 (4.6s) 07:16:08 ~ 07:16:13 (4.7s) 07:14:08 ~ 07:14:13 (4.5s) 07:15:28 ~ 07:15:43 (10.0s)

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。请只输出以下 3 行内容（label 小写，值小写）：
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) 2,510 ms 2,299 ms 3,627 ms 1,783 ms 1,188 ms

首文本(SSE delta) 2,511 ms 2,300 ms 3,628 ms 1,784 ms 8,961 ms

总耗时 5,420 ms 3,777 ms 3,628 ms 3,457 ms 9,170 ms

输入 10 33,555 0 10 10

输出 221 191 146 226 788

缓存创建

237

5m: 0 / 1h: 237

5m: 0 / 1h: 0

221

5m: 0 / 1h: 221

223

5m: 0 / 1h: 223

缓存读取 33,122 0 0 33,122 32,860

输出速度 76.0 tok/s 129.3 tok/s burst 135.1 tok/s burst

响应

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

I don't have explicit confirmation of my specific model identifier in my system context, and I cannot provide a value I'm not certain about. I can confirm:

vendor: anthropic
brand: claude
model: [not

步骤 3: cutoff text

时间 07:16:43 ~ 07:16:49 (4.4s) 07:16:33 ~ 07:16:39 (3.8s) 07:19:53 ~ 07:19:59 (3.8s) 07:16:57 ~ 07:17:03 (3.2s) 07:18:14 ~ 07:18:20 (2.9s)

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) 2,566 ms 2,258 ms 2,583 ms 1,810 ms 1,601 ms

首文本(SSE delta) 2,567 ms 2,259 ms 2,584 ms 1,812 ms 2,118 ms

总耗时 3,478 ms 3,060 ms 2,584 ms 2,282 ms 2,169 ms

输入 10 34,100 0 10 10

输出 69 55 39 66 88

缓存创建

263

5m: 0 / 1h: 263

5m: 0 / 1h: 0

268

5m: 0 / 1h: 268

830

5m: 0 / 1h: 830

缓存读取 33,359 0 0 33,343 33,083

输出速度 75.7 tok/s 68.7 tok/s burst 140.4 tok/s -

响应

2025-02

2025-02

2025-02

2025-02

2024-04

步骤 4: identity_free text

时间 07:19:18 ~ 07:19:28 (7.5s) 07:20:04 ~ 07:20:14 (9.3s) 07:23:46 ~ 07:23:51 (3.4s) 07:19:49 ~ 07:19:59 (5.3s) 07:21:10 ~ 07:21:15 (3.5s)

提示词

查看本步 Prompt

中文回答：你叫什么、属于哪个产品、谁开发？50 字内。

HTTP 头(TTFB) 2,519 ms 2,102 ms 2,347 ms 2,043 ms 1,054 ms

首文本(SSE delta) 2,520 ms 2,103 ms 2,348 ms 2,044 ms 2,153 ms

总耗时 6,519 ms 8,193 ms 2,348 ms 4,297 ms 2,726 ms

输入 10 34,513 0 10 10

输出 350 1,547 30 423 213

缓存创建

109

5m: 0 / 1h: 109

5m: 0 / 1h: 0

104

5m: 0 / 1h: 104

122

5m: 0 / 1h: 122

缓存读取 33,622 0 0 33,611 33,913

输出速度 87.5 tok/s burst burst 187.7 tok/s burst

响应

我叫Claude，由Anthropic公司开发。属于Claude AI助手产品系列，基于Claude Haiku 4.5模型运行。主要用于协助软件工程任务、代码分析和编程工作。

我是Claude Haiku 4.5，由Anthropic开发的AI助手，擅长编程、代码分析和多语言对话。

我是Claude，由Anthropic打造的AI助手。

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手模型。是Claude系列的轻量级版本，设计用于快速、高效的对话和任务处理。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户完成编程、分析、写作等各种任务。

步骤 5: knowledge_recall text

时间 07:21:37 ~ 07:22:32 (51.9s) 07:22:58 ~ 07:23:13 (11.9s) 07:27:12 ~ 07:27:22 (6.5s) 07:23:05 ~ 07:23:40 (30.7s) 07:22:39 ~ 07:23:39 (58.3s)

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) 2,083 ms 1,685 ms 5,119 ms 1,376 ms 13,909 ms

首文本(SSE delta) 2,084 ms 1,686 ms 5,120 ms 1,378 ms 45,408 ms

总耗时 50,823 ms 10,937 ms 5,308 ms 29,546 ms 57,650 ms

输入 10 11,407 0 10 10

输出 4,322 1,740 1,212 4,404 1,157

缓存创建

8,699

5m: 0 / 1h: 8,699

5m: 0 / 1h: 0

8,756

5m: 0 / 1h: 8,756

8,866

5m: 0 / 1h: 8,866

缓存读取 0 0 0 0 0

输出速度 88.7 tok/s 188.1 tok/s burst 156.3 tok/s 94.5 tok/s

响应

```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

{"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

```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":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_pe

步骤 6: digit_count text

时间 07:25:46 ~ 07:26:11 (24.8s) 07:25:11 ~ 07:25:31 (15.3s) 07:28:37 ~ 07:28:47 (8.2s) 07:27:31 ~ 07:27:46 (14.5s) 07:25:30 ~ 07:25:45 (12.0s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=682345261058590859196812636625616149009863990321852060134234752174159751

HTTP 头(TTFB) 1,143 ms 1,113 ms 6,896 ms 1,110 ms 1,278 ms

首文本(SSE delta) 1,144 ms 1,284 ms 6,897 ms 1,112 ms 1,279 ms

总耗时 23,885 ms 14,311 ms 7,093 ms 13,478 ms 11,286 ms

输入 193 208 0 193 139

输出 4,334 2,771 1,305 2,227 2,090

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 190.6 tok/s 212.7 tok/s burst 180.1 tok/s 208.9 tok/s

响应

```json
{"n":72,"c":[7,3,8]}
```

```json
{"n":72,"c":[7,3,8]}
```

{"n":72,"c":[7,3,8]}

```json
{"n":72,"c":[7,3,8]}
```

```json
{"n": 72, "c": [7, 3, 8]}
```

总计

总输入(含缓存) 142,776 146,802 0 142,790 142,946

总输出 9,353 6,348 2,751 7,387 4,379

总缓存创建

42,430

5m: 0 / 1h: 42,430

5m: 0 / 1h: 0

42,471

5m: 0 / 1h: 42,471

42,901

5m: 0 / 1h: 42,901

总缓存读取 100,103 0 0 100,076 99,856

输出速度 116.9 tok/s 190.3 tok/s burst 163.4 tok/s 145.9 tok/s

总执行时间 1m 40.1s 49.6s 30.7s 1m 1.9s 1m 30.1s

总墙钟时间

13m 52.1s

含步骤间等待

13m 14.1s

含步骤间等待

16m 26.0s

含步骤间等待

15m 21.0s

含步骤间等待

13m 22.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1101

Anthropic: Claude Haiku 4.5

$0.1785

Anthropic: Claude Haiku 4.5

>=$0.0138

Anthropic: Claude Haiku 4.5 (成本不完整)

$0.1003

Anthropic: Claude Haiku 4.5

$0.0857

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

ID 格式 msg_01DX47Py7ykrmCsbKPVdFWA1 msg_01dNRFFaGa5mIwLaeWJ5f7C0 msg_2a600c19ddc45f489654d882 msg_01RAYGyMqU74S1dGq9LehZgb msg_018GFRspAeL5kM6xkFwMuzmi

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 set-cookie: server_name_session=7d88fca8b1426ca76c58d5aeba6fb224; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606172325476249749908268d9d6jyEe5pYC	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-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
R-ultra	cf-ray: a0d5cfe9cc59e045-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606172325127688581798268d9d6Fqc6V4gG	id_format: msg_prefix 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_fieldsupstream:anthropic-direct-from-body	a0d5cfe9cc59e045-NRT
u-cc	server: nginx x-new-api-version: 1.0.14 x-oneapi-request-id: 202606172328386588149648268d9d65fwAwoZC	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-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
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_011Cc9hUSWpCGJnWeTU5izZh
O-Max-L2	cf-ray: a0d5d0603e0b8e47-LAX server: Photon-Edge set-cookie: _cfuvid=GXgJZVXEYIuO1_SD_tDB3pETCeu3rgSDzlqxxQEDCyE-1781738731.5573337-1.0.1.1-dKkcaWk2tgjpr8X5tcCZytrox.Ep1hZcI1Sp1_qTGAE; 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	1781738731519-fvzke91nyuau4 req_011Cc9hKWXAW7hqNx5EkA2rT a0d5d0603e0b8e47-LAX

协议指纹评分（机器自动）

总分 / 100 98 69 62 100 83

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 0.0 /10 ×20 0.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

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

sliding 5m 10.0 /10 ×13 5.0 /10 ×13 5.0 /10 ×13 10.0 /10 ×13 10.0 /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 10.0 /10 ×8 10.0 /10 ×8 未计入 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 0.0 /10 ×7

知识截止 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 0.0 /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 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 0.0 /10 ×4 0.0 /10 ×4 10.0 /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

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 98 v3.22.0 69 v3.22.0 62 v3.22.0 基准（参考） 83 v3.22.0

雷达图例

共 23 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 信封自报一致 envelope_self_report_match
21 中文遵循 instruction_following_lang
22 档位思考量 tier_thinking_volume_match
23 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 15644 ms 7250 ms 3951 ms 9252 ms 14254 ms

输出速度 116.9 tok/s 190.3 tok/s burst 163.4 tok/s 145.9 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 ⓘ

$0.1101

$0.1785

>=$0.0138 (成本不完整)

$0.1003

$0.0857

缓存读取占比 ⓘ 70% 0% - 70% 70%

揭晓

通道

服务商 LinkAPI ↗

通道O-Max

服务商 FoxCode ↗

通道R-ultra

服务商 HongMaCC ↗

通道u-cc

官方基准

服务商 SSSAiCode ↗

通道O-Max-L2

返回列表