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

1 任务失败执行不完整

已揭晓揭晓进度

步骤 / 指标

Aiide

O-Max

SudoCode

M-Max

官方基准

FoxCode

R-ultra

TopRouterCN

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 17:29:44 ~ 17:29:54 (5.5s) 17:29:42 ~ 17:29:52 (5.1s) 17:29:48 ~ 17:29:53 (4.3s) 17:29:40 ~ 17:29:45 (4.3s) 17:29:46 ~ 17:29:51 (4.9s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 4,006 ms 3,115 ms 2,851 ms 2,392 ms 3,800 ms

首文本(SSE delta) 4,045 ms 3,116 ms 2,852 ms 2,393 ms 3,801 ms

总耗时 4,299 ms 3,642 ms 3,141 ms 3,313 ms 3,872 ms

输入 10 10 10 33,020 1,645

输出 43 58 53 78 51

缓存创建

33,261

5m: 0 / 1h: 33,261

33,122

5m: 0 / 1h: 33,122

5m: 0 / 1h: 0

31,487

5m: 0 / 1h: 31,487

缓存读取 0 0 33,122 0 0

输出速度 169.3 tok/s 110.3 tok/s 183.4 tok/s 84.8 tok/s -

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 17:32:28 ~ 17:32:38 (6.5s) 17:31:32 ~ 17:31:42 (5.5s) 17:33:52 ~ 17:34:02 (5.6s) 17:32:56 ~ 17:33:06 (5.7s) 17:31:18 ~ 17:31:28 (8.9s)

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) 3,617 ms 2,665 ms 2,773 ms 1,957 ms 7,600 ms

首文本(SSE delta) 3,660 ms 2,666 ms 2,786 ms 1,959 ms 7,601 ms

总耗时 5,464 ms 4,394 ms 4,639 ms 4,803 ms 8,175 ms

输入 10 10 10 33,608 10

输出 205 206 236 338 259

缓存创建

236

5m: 0 / 1h: 236

251

5m: 0 / 1h: 251

246

5m: 0 / 1h: 246

5m: 0 / 1h: 0

1,879

5m: 0 / 1h: 1,879

缓存读取 33,261 33,122 33,122 0 31,487

输出速度 113.6 tok/s 119.2 tok/s 127.4 tok/s 118.8 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 17:35:11 ~ 17:35:17 (4.8s) 17:35:30 ~ 17:35:35 (3.9s) 17:35:52 ~ 17:36:03 (5.0s) 17:35:55 ~ 17:36:00 (4.2s) 17:32:37 ~ 17:32:42 (4.7s)

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) 3,249 ms 2,114 ms 3,447 ms 2,413 ms 3,688 ms

首文本(SSE delta) 3,323 ms 2,116 ms 3,448 ms 2,414 ms 3,690 ms

总耗时 3,633 ms 2,737 ms 3,987 ms 3,076 ms 3,794 ms

输入 10 10 10 34,292 10

输出 64 67 64 52 71

缓存创建

245

5m: 0 / 1h: 245

246

5m: 0 / 1h: 246

259

5m: 0 / 1h: 259

5m: 0 / 1h: 0

282

5m: 0 / 1h: 282

缓存读取 33,497 33,373 33,368 0 33,366

输出速度 206.5 tok/s 107.9 tok/s 118.7 tok/s 78.5 tok/s -

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 17:36:57 ~ 17:37:07 (6.8s) 17:37:56 ~ 17:38:06 (5.7s) 17:38:23 ~ 17:38:33 (5.6s) 17:37:23 ~ 17:37:28 (2.0s) 17:36:17 ~ 17:36:32 (10.6s)

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) 4,119 ms 2,487 ms 2,932 ms - 8,545 ms

首文本(SSE delta) 4,156 ms 2,488 ms 2,934 ms - 8,546 ms

总耗时 5,722 ms 4,690 ms 4,568 ms - 9,720 ms

输入 10 10 10 - 10

输出 250 411 231 - 350

缓存创建

5m: 0 / 1h: 98

105

5m: 0 / 1h: 105

106

5m: 0 / 1h: 106

111

5m: 0 / 1h: 111

缓存读取 33,742 33,619 33,627 - 33,648

输出速度 159.6 tok/s 186.6 tok/s 141.4 tok/s - -

响应

我是Claude，由Anthropic公司开发的AI助手。

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代码编写、问题解答、分析等任务。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行软件工程任务，包括编写代码、调试程序、分析问题等。我致力于提供准确、有帮助且安全的协助。

失败

评测失败

我叫Claude Haiku 4.5，是Claude系列的一个模型。我由Anthropic公司开发。Claude是一个AI助手产品，Haiku是其中最轻量级的版本。

步骤 5: knowledge_recall text

时间 17:39:23 ~ 17:39:43 (16.4s) 17:40:50 ~ 17:41:20 (25.8s) 17:40:33 ~ 17:40:48 (12.3s) - 17:40:15 ~ 17:41:10 (52.6s)

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) 4,412 ms 2,112 ms 1,906 ms - 7,463 ms

首文本(SSE delta) 4,582 ms 2,113 ms 1,908 ms - 7,464 ms

总耗时 15,330 ms 24,550 ms 11,171 ms - 51,794 ms

输入 10 10 10 - 10

输出 1,387 3,334 1,573 - 3,105

缓存创建

8,569

5m: 0 / 1h: 8,569

8,753

5m: 0 / 1h: 8,753

2,861

5m: 0 / 1h: 2,861

8,727

5m: 0 / 1h: 8,727

缓存读取 0 0 5,721 - 0

输出速度 129.0 tok/s 148.6 tok/s 169.8 tok/s - 70.0 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

未执行

```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 17:42:59 ~ 17:43:39 (38.1s) 17:42:36 ~ 17:43:06 (26.5s) 17:42:10 ~ 17:42:35 (23.3s) - 17:43:33 ~ 17:43:58 (22.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=637566465628797062401667027101166561897410778066881806919602992514513180

HTTP 头(TTFB) 1,683 ms 1,809 ms 1,268 ms - 6,601 ms

首文本(SSE delta) 1,684 ms 1,810 ms 1,271 ms - 6,602 ms

总耗时 36,800 ms 25,514 ms 22,357 ms - 21,960 ms

输入 193 193 193 - 193

输出 4,428 4,909 3,235 - 3,746

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 - 0

输出速度 126.1 tok/s 207.1 tok/s 153.4 tok/s - 243.9 tok/s

响应

```json
{"n":72,"c":[9,8,6]}
```

```json
{"n":72,"c":[9,8,6]}
```

```json
{"n":72,"c":[9,8,6]}
```

未执行

```json
{"n":72,"c":[9,8,6]}
```

总计

总输入(含缓存)

143,152

142,834

142,675

100,920

仅前 4 步

142,865

总输出

6,377

8,985

5,392

468

仅前 4 步

7,582

总缓存创建

42,409

5m: 0 / 1h: 42,409

42,477

5m: 0 / 1h: 42,477

3,472

5m: 0 / 1h: 3,472

5m: 0 / 1h: 0

仅前 4 步

42,486

5m: 0 / 1h: 42,486

总缓存读取

100,500

100,114

138,960

仅前 4 步

98,501

输出速度

128.1 tok/s

175.4 tok/s

155.6 tok/s

105.7 tok/s

仅前 4 步

114.8 tok/s

总执行时间

1m 18.2s

1m 12.5s

56.2s

16.2s

仅前 4 步

1m 44.3s

总墙钟时间

13m 55.1s

含步骤间等待

13m 24.0s

含步骤间等待

12m 47.0s

含步骤间等待

7m 47.6s

含步骤间等待

仅前 4 步

14m 12.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0952

Anthropic: Claude Haiku 4.5

$0.1083

Anthropic: Claude Haiku 4.5

$0.0454

Anthropic: Claude Haiku 4.5

$0.1033

Anthropic: Claude Haiku 4.5 (部分)

$0.1027

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count cutoff digit_count

平台 one-api (100%) one-api (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%) Anthropic 直连 (100%)

CDN cloudflare - - cloudflare -

ID 格式 msg_01WnjWfYE4efNjYt96oUTWwC msg_01B7fGtC2iKwqYc1AQTmU9Qz msg_018i5c6rUxc6CRCGf5rAQUgw msg_016Fh4PH2tCfCK6e5YokY66R msg_01DyLHCCYNApRNvuXCsYRjdz

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	cf-ray: a11b43e73d7ffea7-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 202606260943012500195208268d9d61rUhNS2r	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	a11b43e73d7ffea7-AMS
M-Max	server: nginx x-new-api-version: 0b8be5cf x-oneapi-request-id: 20260626174237577032633guEXgEe8	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_011CcRepGc2S3ogkKCzjbfRg
R-ultra	cf-ray: a11b398dff69e041-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606260935576015092008268d9d6p4KJ0f1V	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	a11b398dff69e041-NRT
O-Max	server: nginx/1.27.5	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	tr-ew7y4grd-k7ftlc0abs-052220

协议指纹评分（机器自动）

总分 / 100 98 98 100 97

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 10.0 /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

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 10.0 /10 ×8 10.0 /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

知识截止 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

service_tier 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

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

Req-ID 透传 0.0 /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

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 98 v3.25.0 98 v3.25.0 基准（参考）执行失败，不评分 97 v3.25.0

雷达图例

共 22 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 档位思考量 tier_thinking_volume_match
22 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 11875 ms 10921 ms 8311 ms 3731 ms 16553 ms

输出速度 128.1 tok/s 175.4 tok/s 155.6 tok/s 105.7 tok/s 114.8 tok/s

完成步骤 6/6 成功 6/6 成功 6/6 成功 3/6 成功 · 1 失败 · 2 跳过 6/6 成功

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0952

$0.1083

$0.0454

$0.1033 (部分)

$0.1027

缓存读取占比 ⓘ 70% 70% 97% 0% (部分) 69%

揭晓

通道

服务商 Aiide ↗

通道O-Max

服务商 SudoCode ↗

通道M-Max

官方基准

服务商 FoxCode ↗

通道R-ultra

服务商 TopRouterCN ↗

通道O-Max

返回列表