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

2026/07/02 运行日期

claude-haiku-4-5-20251001 官方基准

2 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

SSSAiCode / R-Kiro
```
评测失败
```

步骤 / 指标

InfAI

R-Kiro

PackyCode

cc-Zeus

官方基准

SSSAiCode

O-Max-L2

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 v2.1.195 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 19:20:13 ~ 19:20:18 (3.6s) 19:20:07 ~ 19:20:17 (5.0s) 19:20:15 ~ 19:20:20 (3.5s) 19:20:11 ~ 19:20:16 (3.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,403 ms 1,355 ms 1,878 ms 2,284 ms

首文本(SSE delta) 2,404 ms 1,357 ms 1,879 ms 2,597 ms

总耗时 2,404 ms 2,053 ms 2,306 ms 2,841 ms

输入 4 10 10 10

输出 1 54 41 54

缓存创建

20,626

5m: 0 / 1h: 20,626

6,166

5m: 0 / 1h: 6,166

6,166

5m: 0 / 1h: 6,166

21,183

5m: 0 / 1h: 21,183

缓存读取 0 15,379 15,379 0

输出速度 - 77.6 tok/s 96.0 tok/s 221.3 tok/s

响应

pong

pong

pong

pong

步骤 2: identity text

时间 19:23:00 ~ 19:23:05 (4.4s) 19:22:16 ~ 19:22:21 (4.3s) 19:21:47 ~ 19:21:57 (6.6s) 19:21:35 ~ 19:21:50 (10.6s)

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。自报身份，输出严格遵守以下 3 行格式（不允许 markdown、代码块、额外文本）：
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) 3,145 ms 1,377 ms 3,235 ms 1,070 ms

首文本(SSE delta) 3,146 ms 1,378 ms 3,236 ms 9,614 ms

总耗时 3,175 ms 3,299 ms 5,590 ms 9,915 ms

输入 12 10 10 10

输出 16 209 239 809

缓存创建

129

5m: 0 / 1h: 129

248

5m: 0 / 1h: 248

235

5m: 0 / 1h: 235

248

5m: 0 / 1h: 248

缓存读取 20,626 21,545 21,545 21,183

输出速度 - 108.8 tok/s 101.5 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-3-5-haiku-20241022

步骤 3: cutoff text

时间 19:27:04 ~ 19:27:09 (4.5s) 19:25:54 ~ 19:25:59 (3.0s) 19:24:50 ~ 19:24:55 (4.0s) 19:23:38 ~ 19:23:43 (3.7s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,992 ms 1,521 ms 2,366 ms 2,396 ms

首文本(SSE delta) 2,993 ms 1,522 ms 2,367 ms 2,719 ms

总耗时 2,994 ms 2,002 ms 3,078 ms 2,964 ms

输入 20 10 10 10

输出 2 63 73 54

缓存创建

5m: 0 / 1h: 28

232

5m: 0 / 1h: 232

279

5m: 0 / 1h: 279

832

5m: 0 / 1h: 832

缓存读取 20,755 21,793 21,780 21,431

输出速度 - 131.3 tok/s 102.7 tok/s 220.4 tok/s

响应

2025-02

2025-02

2025-02

2024-04

步骤 4: identity_free text

时间 19:29:44 ~ 19:29:54 (7.2s) 19:29:07 ~ 19:29:17 (5.2s) 19:28:07 ~ 19:28:17 (7.1s) 19:26:05 ~ 19:26:15 (8.1s)

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) 2,723 ms 2,476 ms 2,499 ms 1,157 ms

首文本(SSE delta) 2,724 ms 2,827 ms 2,500 ms 7,227 ms

总耗时 3,298 ms 4,216 ms 6,171 ms 7,386 ms

输入 28 10 10 10

输出 45 233 472 914

缓存创建

5m: 0 / 1h: 21

5m: 0 / 1h: 97

115

5m: 0 / 1h: 115

5m: 0 / 1h: 96

缓存读取 20,783 22,025 22,059 22,263

输出速度 78.4 tok/s 167.7 tok/s 128.6 tok/s burst

响应

我是由 Anthropic 公司开发的 Claude AI 助手。我是一个大型语言模型，设计用于对话、分析和编程协助。目前运行的是 Claude Haiku 4.5 版本。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助进行编程、分析、写作等各种任务。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。我可以帮助编写代码、分析文本、回答问题、进行研究等多种任务。

我是Claude，Anthropic开发的AI助手。可帮助编程、分析、写作等多种任务。

步骤 5: knowledge_recall text

时间 19:33:43 ~ 19:33:53 (7.8s) 19:31:52 ~ 19:32:32 (38.8s) 19:31:29 ~ 19:31:54 (23.3s) 19:27:21 ~ 19:27:36 (13.0s)

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) 3,083 ms 1,534 ms 1,700 ms 1,368 ms

首文本(SSE delta) 3,084 ms 1,535 ms 1,701 ms 11,966 ms

总耗时 4,630 ms 37,692 ms 22,258 ms 12,261 ms

输入 36 10 10 10

输出 131 6,507 2,521 1,926

缓存创建

9,064

5m: 0 / 1h: 9,064

9,306

5m: 0 / 1h: 9,306

9,597

5m: 0 / 1h: 9,597

10,224

5m: 0 / 1h: 10,224

缓存读取 0 0 0 0

输出速度 84.7 tok/s 180.0 tok/s 122.6 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

```json
{
  "facts": {
    "super_bowl_lviii": "Kansas City Chiefs",
    "oscars2024_best_picture": "Oppenheimer",
    "ucl_2024": "Real Madrid",
    "nba_2024": "Boston Celtics",
    "copa_america_20

```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 19:36:04 ~ 19:36:09 (2.9s) 19:35:52 ~ 19:36:22 (27.7s) 19:33:39 ~ 19:34:09 (27.8s) 19:30:36 ~ 19:31:06 (26.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=909494111497570952351953943543511992391424340995972024281780698660646307

HTTP 头(TTFB) 1,690 ms 1,106 ms 1,442 ms 1,454 ms

首文本(SSE delta) 1,691 ms 1,107 ms 1,443 ms 1,456 ms

总耗时 1,702 ms 26,676 ms 26,824 ms 25,543 ms

输入 145 193 193 139

输出 9 5,562 3,083 4,781

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0

输出速度 - 217.5 tok/s 121.5 tok/s 198.5 tok/s

响应

```json
{"n":66,"c":[14,9,8]}
```

```json
{"n":72,"c":[7,5,14]}
```

```json
{"n":72,"c":[7,5,14]}
```

```json
{"n":72,"c":[7,5,14]}
```

总计

总输入(含缓存) 92,277 97,034 97,398 97,649

总输出 204 12,628 6,429 8,538

总缓存创建

29,868

5m: 0 / 1h: 29,868

16,049

5m: 0 / 1h: 16,049

16,392

5m: 0 / 1h: 16,392

32,583

5m: 0 / 1h: 32,583

总缓存读取 62,164 80,742 80,763 64,877

输出速度 83.0 tok/s 190.7 tok/s 121.1 tok/s 198.9 tok/s

总执行时间 30.4s 1m 24.1s 1m 12.3s 1m 5.1s

总墙钟时间

15m 55.9s

含步骤间等待

16m 15.0s

含步骤间等待

13m 54.2s

含步骤间等待

10m 55.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0448

Anthropic: Claude Haiku 4.5

$0.0915

Anthropic: Claude Haiku 4.5

$0.0610

Anthropic: Claude Haiku 4.5

$0.0901

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%) 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%)

CDN - cloudflare - -

ID 格式 msg_4d86ef76123c4b45b49b2ce7b8aa2b70 msg_017ifPRz9tT2s2aghQHgASdF msg_01M8FkVj6MrcoXFnuJyyNntB msg_01PqzYHnCP9H3Wg47yYYas78

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	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: true usage_has_cache_fields: true	body:id:msg_prefixbody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	202607021136056178490808268d9d6vuvf4Rel
cc-Zeus	via: 1.1 Caddy cf-ray: a14d597edcd0b6a4-LAX server: cloudflare x-oneapi-request-id: 20260702193553435759510MNv4rdX1	id_format: msg_prefix via_chain: ["1.1 Caddy"] 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-rayhdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14d597edcd0b6a4-LAX
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_011CcdAAqdN5xuXJAKrm3mYu
O-Max-L2	cf-ray: a14d51c73d89d908-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	1782991837204-bg01cq31b6e2k req_011Ccd9wLEvTwe9a5NN7ntu9 a14d51c73d89d908-LAX

协议指纹评分（机器自动）

总分 / 100 68 97 100 82

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 6.8 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 6.5 /10 ×20

缓存 TTL 一致 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 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 0.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

系统提示纯净 0.0 /10 ×8 10.0 /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 6.7 /10 ×7

身份(自由) 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

知识截止 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

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

Req-ID 透传 0.0 /10 ×2 0.0 /10 ×2 10.0 /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

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 68 v3.26.0 97 v3.26.0 基准（参考） 82 v3.26.0

雷达图例

共 21 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 信封自报一致 envelope_self_report_match
20 档位思考量 tier_thinking_volume_match
21 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 3034 ms 12656 ms 11038 ms 10152 ms

输出速度 83.0 tok/s 190.7 tok/s 121.1 tok/s 198.9 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 ⓘ

$0.0448

$0.0915

$0.0610

$0.0901

缓存读取占比 ⓘ 67% 83% 83% 66%

揭晓

通道

服务商 InfAI ↗

通道R-Kiro

服务商 PackyCode ↗

通道cc-Zeus

官方基准

服务商 SSSAiCode ↗

通道O-Max-L2

返回列表