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

2026/06/21 运行日期

claude-haiku-4-5-20251001 官方基准

1 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

FastCode

Aiide

O-Max

ByteCatCode

O-Max

ClaudeAPI

O-Bdrk

官方基准

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 17:54:04 ~ 17:54:14 (7.8s) 17:54:04 ~ 17:54:14 (8.0s) 17:54:08 ~ 17:54:18 (5.6s) 17:54:06 ~ 17:54:16 (6.7s) 17:54:10 ~ 17:54:15 (4.4s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,277 ms 4,343 ms 2,673 ms 3,742 ms 2,742 ms

首文本(SSE delta) 3,279 ms 4,346 ms - 3,743 ms 2,744 ms

总耗时 3,911 ms 4,380 ms 3,045 ms 4,134 ms 3,114 ms

输入 10 10 10 10 10

输出 53 52 56 42 44

缓存创建

33,122

5m: 0 / 1h: 33,122

33,261

5m: 0 / 1h: 33,261

33,571

5m: 0 / 1h: 33,571

33,122

5m: 0 / 1h: 33,122

33,122

5m: 0 / 1h: 33,122

缓存读取 0 0 0 0 0

输出速度 83.9 tok/s burst 150.5 tok/s 107.4 tok/s 118.9 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 17:57:49 ~ 17:58:00 (6.5s) 17:55:16 ~ 17:55:27 (7.7s) 17:58:15 ~ 17:58:25 (7.6s) 17:56:45 ~ 17:56:55 (7.0s) 17:56:56 ~ 17:57:01 (4.6s)

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) 3,415 ms 4,748 ms 3,374 ms 3,657 ms 1,762 ms

首文本(SSE delta) 3,416 ms 4,761 ms - 3,659 ms 1,763 ms

总耗时 5,011 ms 6,368 ms 6,285 ms 5,716 ms 3,552 ms

输入 10 10 10 10 10

输出 182 217 227 222 201

缓存创建

246

5m: 0 / 1h: 246

245

5m: 0 / 1h: 245

249

5m: 0 / 1h: 249

33,357

5m: 33,357 / 1h: 0

237

5m: 0 / 1h: 237

缓存读取 33,122 33,261 33,571 0 33,122

输出速度 114.1 tok/s 135.0 tok/s 78.0 tok/s 107.9 tok/s 112.4 tok/s

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

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

步骤 3: cutoff text

时间 17:59:36 ~ 17:59:46 (8.8s) 17:56:31 ~ 17:56:41 (6.3s) 18:01:32 ~ 18:01:37 (4.2s) 17:58:27 ~ 17:58:42 (10.0s) 18:00:15 ~ 18:00:20 (3.5s)

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) 5,887 ms 4,391 ms 2,471 ms 7,989 ms 1,789 ms

首文本(SSE delta) 5,888 ms 4,392 ms - 8,545 ms 1,791 ms

总耗时 7,401 ms 4,877 ms 3,058 ms 8,651 ms 2,342 ms

输入 10 10 10 10 10

输出 85 72 64 66 74

缓存创建

224

5m: 0 / 1h: 224

259

5m: 0 / 1h: 259

250

5m: 0 / 1h: 250

10,086

241

5m: 0 / 1h: 241

缓存读取 33,368 33,506 33,820 23,533 33,359

输出速度 56.2 tok/s 148.5 tok/s 109.0 tok/s - 134.3 tok/s

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 18:02:34 ~ 18:02:49 (11.6s) 18:00:11 ~ 18:00:26 (12.7s) 18:03:41 ~ 18:03:51 (6.1s) 18:02:02 ~ 18:02:12 (8.5s) 18:02:28 ~ 18:02:38 (4.9s)

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) 6,122 ms 3,688 ms 2,990 ms 3,499 ms 2,150 ms

首文本(SSE delta) 6,132 ms 3,947 ms - 3,500 ms 2,152 ms

总耗时 10,336 ms 11,613 ms 5,003 ms 6,916 ms 3,598 ms

输入 10 10 10 10 10

输出 377 979 187 340 185

缓存创建

123

5m: 0 / 1h: 123

114

5m: 0 / 1h: 114

102

5m: 0 / 1h: 102

106

116

5m: 0 / 1h: 116

缓存读取 33,592 33,765 34,070 33,619 33,600

输出速度 89.7 tok/s 127.7 tok/s 92.9 tok/s 99.5 tok/s 127.9 tok/s

响应

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手。Haiku是Claude系列中轻量级但功能强大的模型，适合快速、高效的任务处理。

我是Claude，Anthropic开发的AI助手。可帮助代码编写、调试等软件工程任务，以及对话和信息查询。

我是 Claude Haiku 4.5，由 Anthropic 公司打造的 AI 助手。我可以帮助用户完成软件工程任务，包括代码编写、调试、测试等。我的知识截止日期是 2025 年 2 月。

我是Claude Haiku 4.5，是Anthropic公司开发的Claude AI助手产品系列中的一个模型。

我是Claude，由Anthropic公司开发的AI助手。我能帮助用户进行软件工程任务、代码编辑、问题解决等工作。当前运行的是Claude Haiku 4.5模型。

步骤 5: knowledge_recall text

时间 18:05:33 ~ 18:06:08 (30.2s) 18:01:59 ~ 18:02:14 (13.6s) 18:05:48 ~ 18:06:13 (22.8s) 18:04:41 ~ 18:05:56 (74.6s) 18:05:07 ~ 18:05:32 (20.1s)

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) 2,140 ms 2,922 ms 2,184 ms 3,137 ms 1,737 ms

首文本(SSE delta) 2,142 ms 2,924 ms - 4,018 ms 1,751 ms

总耗时 28,816 ms 12,524 ms 21,136 ms 73,131 ms 19,116 ms

输入 10 10 10 10 10

输出 4,083 1,490 1,701 5,731 2,965

缓存创建

8,710

5m: 0 / 1h: 8,710

9,337

5m: 0 / 1h: 9,337

8,977

5m: 0 / 1h: 8,977

2,605

8,519

5m: 0 / 1h: 8,519

缓存读取 0 0 0 6,078 0

输出速度 153.1 tok/s 155.2 tok/s 89.8 tok/s 82.9 tok/s 170.7 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

```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 18:07:26 ~ 18:07:46 (16.8s) 18:04:23 ~ 18:05:08 (40.2s) 18:09:22 ~ 18:09:42 (16.1s) 18:07:28 ~ 18:07:53 (24.8s) 18:08:38 ~ 18:09:23 (44.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=646456482502468115063277073360168482504406778078574437823465839219415339

HTTP 头(TTFB) 3,013 ms 1,783 ms 3,209 ms 3,214 ms 1,663 ms

首文本(SSE delta) 3,127 ms 1,784 ms - 3,610 ms 1,665 ms

总耗时 15,606 ms 37,894 ms 14,921 ms 23,359 ms 43,188 ms

输入 193 193 643 193 193

输出 2,586 2,786 1,852 3,954 3,929

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 207.2 tok/s 77.2 tok/s 158.1 tok/s 200.2 tok/s 94.6 tok/s

响应

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

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

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

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

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

总计

总输入(含缓存) 142,750 143,991 145,303 142,749 142,559

总输出 7,366 5,596 4,087 10,355 7,398

总缓存创建

42,425

5m: 0 / 1h: 42,425

43,216

5m: 0 / 1h: 43,216

43,149

5m: 0 / 1h: 43,149

79,276

5m: 33,357 / 1h: 33,122

42,235

5m: 0 / 1h: 42,235

总缓存读取 100,082 100,532 101,461 63,230 100,081

输出速度 156.4 tok/s 99.9 tok/s 111.8 tok/s 108.6 tok/s 117.3 tok/s

总执行时间 1m 21.7s 1m 28.5s 1m 2.4s 2m 11.5s 1m 21.7s

总墙钟时间

13m 42.6s

含步骤间等待

11m 4.4s

含步骤间等待

15m 34.3s

含步骤间等待

13m 47.4s

含步骤间等待

15m 13.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1001

Anthropic: Claude Haiku 4.5

$0.0923

Anthropic: Claude Haiku 4.5

$0.0852

Anthropic: Claude Haiku 4.5

$0.1258

Anthropic: Claude Haiku 4.5

$0.1000

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%) one-api (100%) one-api (100%) one-api (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_014UHn8sbt4nLM8sVpKaqZ6C msg_01JLYoqqoRkjx94bL8oc6sft msg_01L7iYDQdXhv6Ttd6g6MPcvW msg_01CPHfWW6yzQFn7oZyexAPrC msg_01QE3TbS6bNeMYNVwEcA2sVg

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
cc	server: nginx	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	—
O-Max	cf-ray: a0f23067ba770e10-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 202606211004264403746508268d9d6ZsmxRm57	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	a0f23067ba770e10-AMS
O-Max	server: nginx/1.29.8 x-oneapi-request-id: 20260621100924974715208268d9d6lCm0lV7m	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-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
O-Bdrk	x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606211007303088064588268d9d6HgLmbwqS	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_011CcGDnkjD5A6jYWwWEYv7A

协议指纹评分（机器自动）

总分 / 100 98 97 92 80 100

基线状态已对照基线已对照基线已对照基线已对照基线已对照基线

缓存命中比 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 4.4 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 10.0 /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 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 7.7 /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 5.0 /10 ×8 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 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 4.0 /10 ×5 10.0 /10 ×5

延迟基线 9.0 /10 ×5 7.8 /10 ×5 10.0 /10 ×5 7.6 /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

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 v3.24.0 97 v3.24.0 92 v3.24.0 80 v3.24.0 基准（参考）

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 11847 ms 12943 ms 8908 ms 20318 ms 12485 ms

输出速度 156.4 tok/s 99.9 tok/s 111.8 tok/s 108.6 tok/s 117.3 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 ⓘ

$0.1001

$0.0923

$0.0852

$0.1258

$0.1000

缓存读取占比 ⓘ 70% 70% 70% 44% 70%

揭晓

通道

服务商 FastCode ↗

通道cc

服务商 Aiide ↗

通道O-Max

服务商 ByteCatCode ↗

通道O-Max

服务商 ClaudeAPI ↗

通道O-Bdrk

官方基准

返回列表