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

2026/06/25 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

DBAI / M-
```
评测失败
```

步骤 / 指标

TopRouterCN

O-Max

WorldBase.ai

M-

官方基准

Aiide

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 v2.1.160 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 19:31:01 ~ 19:31:06 (4.5s) 19:31:03 ~ 19:31:13 (5.3s) 19:31:07 ~ 19:31:12 (3.5s) 19:30:59 ~ 19:31:09 (7.9s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,227 ms 3,622 ms 2,231 ms 3,861 ms

首文本(SSE delta) 3,228 ms 3,688 ms 2,233 ms 3,863 ms

总耗时 3,307 ms 3,722 ms 2,547 ms 4,468 ms

输入 1,645 4 10 10

输出 37 1 49 46

缓存创建

31,487

5m: 0 / 1h: 31,487

31,129

5m: 0 / 1h: 31,129

33,122

5m: 0 / 1h: 33,122

33,261

5m: 0 / 1h: 33,261

缓存读取 0 0 0 0

输出速度 - - 156.1 tok/s 76.0 tok/s

响应

pong

pong

pong

pong

步骤 2: identity text

时间 19:34:15 ~ 19:34:25 (5.2s) 19:32:58 ~ 19:33:08 (5.0s) 19:33:30 ~ 19:33:35 (4.4s) 19:32:47 ~ 19:32:57 (5.7s)

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。请按下列 3 行格式精确填空作答（label 不变，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) 4,257 ms 3,554 ms 1,594 ms 3,098 ms

首文本(SSE delta) 4,259 ms 3,555 ms 1,596 ms 3,099 ms

总耗时 4,328 ms 3,879 ms 3,371 ms 4,589 ms

输入 10 12 10 10

输出 198 16 217 190

缓存创建

1,856

5m: 0 / 1h: 1,856

122

5m: 0 / 1h: 122

233

5m: 0 / 1h: 233

230

5m: 0 / 1h: 230

缓存读取 31,487 31,129 33,122 33,261

输出速度 burst - 122.3 tok/s 127.5 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

步骤 3: cutoff text

时间 19:37:06 ~ 19:37:11 (4.1s) 19:34:28 ~ 19:34:38 (5.8s) 19:35:52 ~ 19:35:57 (3.3s) 19:34:07 ~ 19:34:17 (5.5s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,226 ms 3,349 ms 1,820 ms 3,773 ms

首文本(SSE delta) 3,227 ms 3,672 ms 1,821 ms 3,775 ms

总耗时 3,328 ms 3,672 ms 2,268 ms 4,312 ms

输入 10 20 10 10

输出 66 2 68 70

缓存创建

221

5m: 0 / 1h: 221

5m: 0 / 1h: 36

240

5m: 0 / 1h: 240

232

5m: 0 / 1h: 232

缓存读取 33,343 31,251 33,355 33,491

输出速度 - - 152.1 tok/s 130.4 tok/s

响应

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 19:39:49 ~ 19:39:59 (8.3s) 19:36:42 ~ 19:36:52 (6.4s) 19:39:09 ~ 19:39:14 (3.7s) 19:36:46 ~ 19:36:56 (9.8s)

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) 7,462 ms 3,883 ms 1,629 ms 3,174 ms

首文本(SSE delta) 7,463 ms 3,930 ms 1,631 ms 3,178 ms

总耗时 7,548 ms 4,389 ms 2,759 ms 5,734 ms

输入 10 28 10 10

输出 651 45 179 378

缓存创建

100

5m: 0 / 1h: 100

5m: 0 / 1h: 17

102

5m: 0 / 1h: 102

112

5m: 0 / 1h: 112

缓存读取 33,564 31,287 33,595 33,723

输出速度 burst 98.0 tok/s 158.7 tok/s 147.9 tok/s

响应

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手，可帮助完成编程、代码审查等软件工程任务。

我是 Kiro，一个由 Anthropic 开发的 AI 代理。我基于 Claude 模型构建，帮助用户完成软件工程任务，包括编写代码、修复 bug、重构代码等。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助编程、研究、写作和各种分析任务。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行编程、代码分析、问题解决和创意写作等任务。

步骤 5: knowledge_recall text

时间 19:43:57 ~ 19:44:27 (29.5s) 19:38:18 ~ 19:38:23 (4.5s) 19:41:11 ~ 19:41:31 (19.5s) 19:38:21 ~ 19:38:41 (19.3s)

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) 6,996 ms 2,335 ms 1,975 ms 2,928 ms

首文本(SSE delta) 6,997 ms 2,348 ms 1,977 ms 2,930 ms

总耗时 28,795 ms 3,460 ms 18,552 ms 18,027 ms

输入 10 36 10 10

输出 4,098 136 2,408 2,534

缓存创建

8,933

5m: 0 / 1h: 8,933

8,323

5m: 0 / 1h: 8,323

8,494

5m: 0 / 1h: 8,494

8,071

5m: 0 / 1h: 8,071

缓存读取 0 0 0 0

输出速度 188.0 tok/s 122.3 tok/s 145.3 tok/s 167.8 tok/s

响应

{"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":"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 19:48:00 ~ 19:48:20 (17.0s) 19:40:08 ~ 19:40:13 (3.6s) 19:45:28 ~ 19:45:53 (20.9s) 19:40:31 ~ 19:40:51 (16.4s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=019927692870765549021878856804477888173855931434219929520354618774866034

HTTP 头(TTFB) 5,986 ms 2,108 ms 1,119 ms 1,524 ms

首文本(SSE delta) 5,987 ms 2,118 ms 1,121 ms 1,525 ms

总耗时 16,268 ms 2,320 ms 19,883 ms 15,300 ms

输入 193 145 193 193

输出 2,563 8 3,611 2,690

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 249.3 tok/s - 192.5 tok/s 195.3 tok/s

响应

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

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

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

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

总计

总输入(含缓存) 142,869 133,539 142,506 142,624

总输出 7,613 208 6,532 5,908

总缓存创建

42,597

5m: 0 / 1h: 42,597

39,627

5m: 0 / 1h: 39,627

42,191

5m: 0 / 1h: 42,191

41,906

5m: 0 / 1h: 41,906

总缓存读取 98,394 93,667 100,072 100,475

输出速度 207.6 tok/s 115.2 tok/s 167.5 tok/s 173.5 tok/s

总执行时间 1m 8.5s 30.6s 55.3s 1m 4.6s

总墙钟时间

17m 19.0s

含步骤间等待

9m 9.9s

含步骤间等待

14m 46.0s

含步骤间等待

9m 52.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1030

Anthropic: Claude Haiku 4.5

$0.0602

Anthropic: Claude Haiku 4.5

$0.0956

Anthropic: Claude Haiku 4.5

$0.0922

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%) 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 - cloudflare

ID 格式 msg_01DLEhvSqDD9rUUfSHYoUURh msg_c6bc3bf6401b422c9606c2528b9e37b7 msg_016fL6ug17ZjMR2vJKBcwPsm msg_01TMrymLJWnA2rGPNU9zTh2b

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/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-o4yoh9bo9h-093777
M-	cf-ray: a113b2216ba00b66-AMS server: cloudflare	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	hdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	65a95373-057b-4d78-b1c1-5f232f13ac76 a113b2216ba00b66-AMS
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_011CcPvQhLFjS3BjxZ6iUVTu
O-Max	cf-ray: a113b2b0cc72f85f-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 20260625114033167275358268d9d6d3a1zBSX	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	a113b2b0cc72f85f-AMS

协议指纹评分（机器自动）

总分 / 100 98 67 100 98

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 0.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

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

service_tier 10.0 /10 ×6 0.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

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

延迟基线 9.3 /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 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 v3.24.1 67 v3.24.1 基准（参考） 98 v3.24.1

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 10596 ms 3574 ms 8230 ms 8738 ms

输出速度 207.6 tok/s 115.2 tok/s 167.5 tok/s 173.5 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 ⓘ

$0.1030

$0.0602

$0.0956

$0.0922

缓存读取占比 ⓘ 69% 70% 70% 70%

揭晓

通道

服务商 TopRouterCN ↗

通道O-Max

服务商 WorldBase.ai ↗

通道M-

官方基准

服务商 Aiide ↗

通道O-Max

返回列表