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-sonnet-4-6 官方基准

2 通道分数 < 70 另含 1 任务失败

已揭晓揭晓进度

步骤 / 指标

RunAPI

O-Max

ClaudeAPI

O-Bdrk

官方基准

ClaudeCN

O-Max

SSSAiCode

R-Kiro

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6 claude-sonnet-4-6

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 12:29:20 ~ 12:29:30 (5.3s) 12:29:18 ~ 12:29:28 (6.2s) 12:29:22 ~ 12:29:27 (4.8s) 12:29:16 ~ 12:29:26 (6.7s) 12:29:14 ~ 12:29:19 (3.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,189 ms 3,622 ms 3,155 ms 4,519 ms 2,124 ms

首文本(SSE delta) 3,190 ms 3,623 ms 3,156 ms 4,520 ms 2,207 ms

总耗时 3,477 ms 3,904 ms 3,263 ms 4,842 ms 2,210 ms

输入 3 3 3 3 0

输出 5 5 5 5 2

缓存创建

32,319

5m: 0 / 1h: 32,319

31,628

5m: 31,628 / 1h: 0

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 0 0 0

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 12:31:08 ~ 12:31:13 (4.4s) 12:33:10 ~ 12:33:20 (5.8s) 12:30:39 ~ 12:30:49 (5.9s) 12:32:18 ~ 12:32:28 (6.7s) 12:33:00 ~ 12:33:05 (3.4s)

提示词

查看本步 Prompt

输出 3 行 metadata 描述你自己（当前 API 模型）。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。禁止任何额外字符：
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) - 3,578 ms 4,007 ms 4,463 ms 2,349 ms

首文本(SSE delta) - 3,579 ms 4,009 ms 4,464 ms 2,364 ms

总耗时 - 3,875 ms 4,797 ms 5,184 ms 2,626 ms

输入 - 3 3 3 0

输出 - 22 22 22 17

缓存创建

189

5m: 189 / 1h: 0

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

缓存读取 - 31,628 32,318 32,318 0

响应

失败

评测失败

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

vendor: anthropic
brand: claude
model: claude-sonnet-4-6

步骤 3: cutoff text

时间 - 12:34:53 ~ 12:35:03 (5.7s) 12:34:06 ~ 12:34:16 (7.2s) 12:35:56 ~ 12:36:06 (5.6s) 12:35:41 ~ 12:35:46 (2.8s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) - 4,064 ms 5,851 ms 4,244 ms 1,679 ms

首文本(SSE delta) - 4,065 ms 5,968 ms 4,245 ms 1,904 ms

总耗时 - 4,263 ms 6,086 ms 4,517 ms 1,908 ms

输入 - 3 3 3 0

输出 - 8 8 8 3

缓存创建

32,545

5m: 0 / 1h: 32,545

5m: 0 / 1h: 55

5m: 0 / 1h: 38

缓存读取 - 0 32,507 32,507 0

响应

未执行

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 - 12:37:50 ~ 12:38:00 (8.2s) 12:37:55 ~ 12:38:00 (4.1s) 12:39:10 ~ 12:39:20 (6.6s) 12:37:52 ~ 12:37:57 (3.6s)

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) - 4,974 ms 1,929 ms 4,281 ms 1,596 ms

首文本(SSE delta) - 4,976 ms 1,931 ms 4,282 ms 1,700 ms

总耗时 - 5,751 ms 2,598 ms 5,257 ms 2,713 ms

输入 - 3 3 3 0

输出 - 59 46 62 36

缓存创建

32,580

5m: 0 / 1h: 32,580

5m: 0 / 1h: 39

5m: 0 / 1h: 43

缓存读取 - 0 32,562 32,545 0

输出速度 - 76.1 tok/s 69.0 tok/s 63.6 tok/s 35.5 tok/s

响应

未执行

我是 Claude，由 Anthropic 开发的 AI 助手。我基于 claude-sonnet-4-6 模型，旨在提供安全、有益的对话和任务协助。

我是 Claude，由 Anthropic 公司打造的 AI 助手。当前运行的模型版本为 claude-sonnet-4-6。

我是 Claude，由 Anthropic 开发的 AI 助手。我基于大型语言模型构建，能够帮助用户完成写作、分析、编程、问答等多种任务。

我是 Kiro，一个由 Anthropic 技术驱动的 AI 开发环境助手，基于 Claude Sonnet 4.6 模型构建，帮助开发者更高效地完成软件工程任务。

步骤 5: knowledge_recall text

时间 - 12:39:17 ~ 12:44:42 (323.3s) 12:40:19 ~ 12:40:29 (7.1s) 12:42:21 ~ 12:42:31 (6.7s) 12:40:40 ~ 12:40:45 (4.3s)

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) - 3,474 ms 3,883 ms 2,764 ms 1,736 ms

首文本(SSE delta) - 3,476 ms 4,000 ms 2,765 ms 1,788 ms

总耗时 - 315,973 ms 5,798 ms 5,257 ms 3,599 ms

输入 - 3 3 3 0

输出 - - 205 205 221

缓存创建

8,028

5m: 0 / 1h: 8,028

8,036

5m: 0 / 1h: 8,036

8,039

5m: 0 / 1h: 8,039

缓存读取 - 0 0 0 0

输出速度 - — (thinking-only) 114.0 tok/s 82.3 tok/s 122.0 tok/s

响应

未执行

{"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

{"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

{"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

{"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

步骤 6: digit_count text

时间 - 12:46:59 ~ 12:47:29 (26.3s) 12:42:10 ~ 12:42:30 (19.0s) 12:45:52 ~ 12:46:17 (22.6s) 12:43:28 ~ 12:43:58 (26.4s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=055598680854757900115316876198321358275270300007682202619274525176906545

HTTP 头(TTFB) - 2,215 ms 1,378 ms 2,277 ms 1,424 ms

首文本(SSE delta) - 2,254 ms 1,380 ms 2,943 ms 1,427 ms

总耗时 - 24,784 ms 18,103 ms 21,434 ms 25,296 ms

输入 - 164 164 164 0

输出 - 1,687 1,870 1,511 1,750

缓存创建

5m: 0 / 1h: 0

缓存读取 - 0 0 0 0

输出速度 - 74.9 tok/s 111.8 tok/s 81.7 tok/s 73.3 tok/s

响应

未执行

```
s = "055598680854757900115316876198321358275270300007682202619274525176906545"
```

Indexing each character:

| # | digit | # | digit | # | digit | # | digit |
|---|-------|---|-------|---|-------

## Counting s = `055598680854757900115316876198321358275270300007682202619274525176906545`

**Splitting into groups of 10 for accuracy:**

| Group | Chars | 0s | 7s | 9s |
|-------|-------|----|----|-

**Counting the string:**

`055598680854757900115316876198321358275270300007682202619274525176906545`

**Total length:** 72

**Digit-by-digit scan:**

| # | char | | # | char | | # | char |
|---|------

```
s = "055598680854757900115316876198321358275270300007682202619274525176906545"
```

Indexing each character (0-based):

| idx | digit | idx | digit | idx | digit | idx
|-----|-------|-----|-------

总计

总输入(含缓存)

32,322

仅前 2 步

136,777

138,203

138,176

总输出

仅前 2 步

1,781

2,156

1,813

2,029

总缓存创建

32,319

5m: 0 / 1h: 32,319

仅前 2 步

104,970

5m: 31,817 / 1h: 73,153

40,637

5m: 0 / 1h: 40,637

40,627

5m: 0 / 1h: 40,627

总缓存读取

仅前 2 步

31,628

97,387

97,370

输出速度

仅前 2 步

74.9 tok/s

110.5 tok/s

81.0 tok/s

75.2 tok/s

总执行时间

9.7s

仅前 2 步

6m 15.5s

48.1s

54.9s

43.5s

总墙钟时间

1m 53.3s

含步骤间等待

仅前 2 步

18m 11.6s

含步骤间等待

13m 8.1s

含步骤间等待

17m 1.0s

含步骤间等待

14m 44.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1213

Anthropic: Claude Sonnet 4.6 (部分)

$0.3856

Anthropic: Claude Sonnet 4.6

$0.2145

Anthropic: Claude Sonnet 4.6

$0.2093

Anthropic: Claude Sonnet 4.6

>=$0.0304

Anthropic: Claude Sonnet 4.6 (成本不完整)

通道指纹

来源步骤 ping digit_count digit_count digit_count digit_count

平台 one-api (100%) one-api (100%) direct (100%) one-api (100%) direct (75%)

上游 Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (75%)

CDN - - - cloudflare -

ID 格式 msg_01NuQqtsv295readrifHrLFv msg_01GjphVA7rGh2n1CidAp5oV8 msg_01SyRtLMaDE19nH9tcBPc7XV msg_01AVew531rF6GbC8fTtSMyhP msg_33346ac9a0c6446c94fb71ba94411708

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/1.24.0 (Ubuntu) x-oneapi-request-id: 20260621122922978467723oFaSvOI8	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	—
O-Bdrk	x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606210447012862669158268d9d69ADpxLdw	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idbody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody: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: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcFntpfr1utSXd1TLHTcy
O-Max	cf-ray: a0f05dc99a350e2f-AMS server: cloudflare x-oneapi-request-id: 20260621124554100117389R17iIodg	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f05dc99a350e2f-AMS
R-Kiro	server: Photon-Edge x-amzn-requestid: 9050579c-bd5f-47a6-9211-d3508c2cbe71	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: true	body:id:msg_prefixbody:usage.cache_fieldsupstream:x-amzn-requestidupstream:anthropic-direct-from-body	1782017009332-xjhsemmcrct81 9050579c-bd5f-47a6-9211-d3508c2cbe71

协议指纹评分（机器自动）

总分 / 100 71 100 98 52

基线状态已对照基线已对照基线已对照基线已对照基线

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 5.0 /10 ×14

sliding 5m — 0.0 /10 ×13 10.0 /10 ×13 10.0 /10 ×13 5.0 /10 ×13

缓存 TTL 一致 — 4.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 0.0 /10 ×15

缓存命中比 — 0.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 0.0 /10 ×20

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 0.0 /10 ×7

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

Req-ID 透传 — 0.0 /10 ×4 10.0 /10 ×4 0.0 /10 ×4 0.0 /10 ×4

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

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

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

SDK 一致 — 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

系统提示纯净 — 5.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8 未计入

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

流式投递 — 未计入 10.0 /10 ×5 10.0 /10 ×5 10.0 /10 ×5

综合结论

与基线相似度执行失败，不评分 71 v3.24.0 基准（参考） 98 v3.24.0 52 v3.24.0

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 缓存连续 cache_continuity_intra
13 知识召回 knowledge_recall_match
14 延迟基线 latency_baseline_match
15 原生 msg-ID anthropic_msg_id_format
16 Sliding 缓存 cache_sliding_correctness
17 结构化身份 identity_structured_match
18 信封自报一致 envelope_self_report_match
19 中文遵循 instruction_following_lang
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 3477 ms 59758 ms 6774 ms 7749 ms 6392 ms

输出速度 - 74.9 tok/s 110.5 tok/s 81.0 tok/s 75.2 tok/s

完成步骤 1/6 成功 · 1 失败 · 4 跳过 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 ⓘ

$0.1213 (部分)

$0.3856

$0.2145

$0.2093

>=$0.0304 (成本不完整)

缓存读取占比 ⓘ 0% (部分) 23% 70% 70% -

揭晓

通道

服务商 RunAPI ↗

通道O-Max

服务商 ClaudeAPI ↗

通道O-Bdrk

官方基准

服务商 ClaudeCN ↗

通道O-Max

服务商 SSSAiCode ↗

通道R-Kiro

返回列表