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

2026/07/01 运行日期

claude-haiku-4-5-20251001 官方基准

2 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

NowCoding

R-Kiro

官方基准

ClaudeZZ

R-kiro

OmniaKey

O-Max

SSSAiCode

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 v2.1.195 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 23:18:17 ~ 23:18:27 (5.5s) 23:18:21 ~ 23:18:26 (4.3s) 23:18:13 ~ 23:18:23 (5.9s) 23:18:20 ~ 23:18:30 (6.5s) 23:18:15 ~ 23:18:20 (3.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,118 ms 1,825 ms 3,464 ms 4,430 ms 1,979 ms

首文本(SSE delta) 2,126 ms 1,827 ms 3,466 ms 4,432 ms 2,308 ms

总耗时 2,656 ms 2,255 ms 4,206 ms 4,433 ms 2,452 ms

输入 21,419 10 10 10 10

输出 68 51 35 50 54

缓存创建

6,166

5m: 0 / 1h: 6,166

13,155

5m: 0 / 1h: 13,155

6,166

5m: 0 / 1h: 6,166

21,177

5m: 0 / 1h: 21,177

缓存读取 0 15,379 0 15,379 0

输出速度 128.3 tok/s 119.2 tok/s 47.3 tok/s burst -

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 23:19:59 ~ 23:20:04 (4.8s) 23:19:48 ~ 23:19:58 (5.5s) 23:20:07 ~ 23:20:17 (6.3s) 23:20:47 ~ 23:20:57 (9.5s) 23:22:19 ~ 23:22:39 (18.1s)

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) 2,201 ms 1,972 ms 2,930 ms 3,731 ms 2,444 ms

首文本(SSE delta) 2,202 ms 1,974 ms 2,932 ms 3,732 ms 16,945 ms

总耗时 3,696 ms 3,939 ms 5,526 ms 6,014 ms 17,273 ms

输入 21,617 10 10 10 10

输出 202 256 218 233 1,324

缓存创建

231

5m: 0 / 1h: 231

222

5m: 0 / 1h: 222

230

5m: 0 / 1h: 230

234

5m: 0 / 1h: 234

缓存读取 0 21,545 13,155 21,545 21,177

输出速度 135.2 tok/s 130.3 tok/s 84.0 tok/s 102.1 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

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

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

步骤 3: cutoff text

时间 23:22:56 ~ 23:23:01 (3.6s) 23:21:49 ~ 23:21:54 (3.2s) 23:24:04 ~ 23:24:14 (6.3s) 23:22:36 ~ 23:22:46 (6.1s) 23:25:04 ~ 23:25:09 (3.4s)

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) 1,862 ms 1,738 ms 4,325 ms 3,570 ms 2,126 ms

首文本(SSE delta) 1,864 ms 1,739 ms 4,327 ms 3,571 ms 2,577 ms

总耗时 2,331 ms 2,251 ms 5,534 ms 4,554 ms 2,614 ms

输入 21,678 10 10 10 10

输出 69 74 69 65 74

缓存创建

298

5m: 0 / 1h: 298

284

5m: 0 / 1h: 284

273

5m: 0 / 1h: 273

1,347

5m: 0 / 1h: 1,347

缓存读取 0 21,776 13,377 21,775 21,411

输出速度 147.8 tok/s 144.5 tok/s 57.2 tok/s 66.1 tok/s burst

响应

2025-02

2025-02

2025-02

2025-02

2024-04

步骤 4: identity_free text

时间 23:26:04 ~ 23:26:09 (4.1s) 23:24:26 ~ 23:24:41 (12.9s) 23:28:08 ~ 23:28:18 (6.5s) 23:26:15 ~ 23:26:25 (8.7s) 23:27:06 ~ 23:27:16 (7.7s)

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) 2,032 ms 1,626 ms 3,350 ms 4,130 ms 2,070 ms

首文本(SSE delta) 2,033 ms 1,669 ms 3,351 ms 4,131 ms 6,499 ms

总耗时 3,046 ms 11,920 ms 5,360 ms 5,500 ms 6,868 ms

输入 21,709 10 10 10 10

输出 120 2,001 190 154 223

缓存创建

116

5m: 0 / 1h: 116

113

5m: 0 / 1h: 113

105

5m: 0 / 1h: 105

114

5m: 0 / 1h: 114

缓存读取 0 22,074 13,661 22,048 22,758

输出速度 118.5 tok/s 195.2 tok/s 94.6 tok/s 112.5 tok/s burst

响应

我是Claude Haiku 4.5，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、代码编写、分析和其他工作。

我是Claude，Anthropic开发的AI助手。帮助代码编写、软件工程等任务。具备代码理解和生成能力，致力提供安全有效的协助。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户完成编程和软件工程任务，提供代码建议、调试支持和技术解决方案。

我叫Claude，是由Anthropic公司开发的AI助手。我属于Claude产品家族，这个系列包括多个不同规模的模型版本，用于提供对话和代码协助等服务。

我叫Claude，属于Claude产品系列，由Anthropic公司开发。

步骤 5: knowledge_recall text

时间 23:28:06 ~ 23:28:21 (12.5s) 23:27:38 ~ 23:27:58 (15.5s) 23:31:58 ~ 23:32:28 (28.0s) 23:29:23 ~ 23:29:48 (20.1s) 23:30:27 ~ 23:30:42 (12.4s)

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) 1,995 ms 3,330 ms 2,789 ms 2,652 ms 3,297 ms

首文本(SSE delta) 1,997 ms 3,332 ms 2,790 ms 2,653 ms 10,731 ms

总耗时 11,585 ms 14,515 ms 27,159 ms 18,558 ms 11,262 ms

输入 9,537 10 10 10 10

输出 1,452 2,224 2,867 2,288 1,071

缓存创建

11,142

5m: 0 / 1h: 11,142

9,388

5m: 0 / 1h: 9,388

9,258

5m: 0 / 1h: 9,258

10,046

5m: 0 / 1h: 10,046

缓存读取 0 0 0 0 0

输出速度 151.4 tok/s 198.9 tok/s 117.6 tok/s 143.9 tok/s burst

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

{"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

{"facts":{"super_bowl_lviii":"Kansas City Chiefs","oscars2024_best_picture":"Oppenheimer","ucl_2024":"UNKNOWN","nba_2024":"UNKNOWN","copa_america_2024":"UNKNOWN","euro2024":"UNKNOWN","nobel_peace_2024

步骤 6: digit_count text

时间 23:31:45 ~ 23:32:00 (10.4s) 23:29:16 ~ 23:29:36 (17.3s) 23:34:35 ~ 23:34:55 (15.0s) 23:32:55 ~ 23:33:56 (58.2s) 23:32:42 ~ 23:33:02 (15.3s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=597514358848772279782479139634089496255731620014388930652178220673608539

HTTP 头(TTFB) 1,154 ms 862 ms 2,382 ms 2,202 ms 1,532 ms

首文本(SSE delta) 1,156 ms 863 ms 2,384 ms 2,203 ms 1,533 ms

总耗时 9,451 ms 15,765 ms 13,869 ms 57,063 ms 14,676 ms

输入 237 193 10 193 139

输出 849 2,443 1,773 7,029 2,469

缓存创建

5m: 0 / 1h: 0

5,980

5m: 0 / 1h: 5,980

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 102.4 tok/s 163.9 tok/s 154.4 tok/s 128.1 tok/s 187.9 tok/s

响应

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

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

I'll count the digits in the string `s = "597514358848772279782479139634089496255731620014388930652178220673608539"`

**Length:** 72 characters

**Digit counts:**
- 0s: 6
- 7s: 9
- 9s: 8

```json
{"n"

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

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

总计

总输入(含缓存) 96,197 98,970 69,395 97,022 98,453

总输出 2,760 7,049 5,152 9,819 5,215

总缓存创建

17,953

5m: 0 / 1h: 17,953

29,142

5m: 0 / 1h: 29,142

16,032

5m: 0 / 1h: 16,032

32,918

5m: 0 / 1h: 32,918

总缓存读取 0 80,774 40,193 80,747 65,346

输出速度 129.1 tok/s 179.6 tok/s 121.5 tok/s 129.6 tok/s 187.9 tok/s

总执行时间 40.9s 58.8s 1m 8.0s 1m 49.1s 1m 0.3s

总墙钟时间

13m 43.1s

含步骤间等待

11m 15.0s

含步骤间等待

16m 41.3s

含步骤间等待

15m 36.1s

含步骤间等待

14m 47.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1100

Anthropic: Claude Haiku 4.5

$0.0660

Anthropic: Claude Haiku 4.5

$0.0663

Anthropic: Claude Haiku 4.5

$0.0775

Anthropic: Claude Haiku 4.5

$0.0739

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%) 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%) Anthropic 直连 (100%)

CDN - - - cloudflare -

ID 格式 msg_a5475bd5d457466fba2b3fb95769eb41 msg_013crhGrqXkxaEbBru1rmsu7 msg_016BGsM1e3fPjAS1HhX1gHrU msg_016Bd9oK4ghf3DqhU9x9pAQU msg_01RMZ5beZuADuNHiLArk7wwz

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
R-Kiro	server: nginx x-new-api-version: v0.0.0 x-oneapi-request-id: 20260701153146873102301lgWrssRo	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:model:date_suffixbody:usage.cache_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_011CcbaL5W5c1jXmZzz91ASb
R-kiro	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: a14677616acefea7-AMS server: cloudflare x-oneapi-request-id: 202607011532576631879028268d9d6D4D94tVG	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-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	a14677616acefea7-AMS
O-Max	cf-ray: a146770ddf90e9e4-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	1782919963712-8m87dskl0hckq req_011CcbabDkbLVNTMpnuqz8XW a146770ddf90e9e4-LAX

协议指纹评分（机器自动）

总分 / 100 55 97 87 96 78

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 0.0 /10 ×20 10.0 /10 ×20 8.8 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 0.0 /10 ×15 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 7.8 /10 ×15 6.7 /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 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 5.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13

原生 msg-ID 0.0 /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 0.0 /10 ×8 10.0 /10 ×8 5.0 /10 ×8

身份(结构) 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 6.7 /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 0.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

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

inference_geo 0.0 /10 ×5 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 10.0 /10 ×3

Req-ID 透传 0.0 /10 ×2 10.0 /10 ×2 0.0 /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 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 55 v3.26.0 基准（参考） 87 v3.26.0 96 v3.26.0 78 v3.26.0

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 5461 ms 8441 ms 10276 ms 16020 ms 9191 ms

输出速度 129.1 tok/s 179.6 tok/s 121.5 tok/s 129.6 tok/s 187.9 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 ⓘ

$0.1100

$0.0660

$0.0663

$0.0775

$0.0739

缓存读取占比 ⓘ 0% 82% 58% 83% 66%

揭晓

通道

服务商 NowCoding ↗

通道R-Kiro

官方基准

服务商 ClaudeZZ ↗

通道R-kiro

服务商 OmniaKey ↗

通道O-Max

服务商 SSSAiCode ↗

通道O-Max

返回列表