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-sonnet-4-6 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

0-0 / O-Max
```
评测失败
```

步骤 / 指标

sucui

O-Max-NL

Aiide

O-Max

官方基准

luminai

M-Kiro/Antigravity

基本信息

服务名 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

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 17:37:43 ~ 17:37:53 (6.0s) 17:37:41 ~ 17:37:51 (7.5s) 17:37:47 ~ 17:37:52 (4.0s) 17:37:45 ~ 17:37:55 (6.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,899 ms 6,197 ms 2,710 ms 3,947 ms

首文本(SSE delta) 4,847 ms 6,380 ms 2,712 ms 3,948 ms

总耗时 4,847 ms 6,395 ms 2,953 ms 3,948 ms

输入 3 3 3 1,334

输出 5 5 5 1

缓存创建

13,122

5m: 0 / 1h: 13,122

20,875

5m: 0 / 1h: 20,875

5,991

5m: 0 / 1h: 5,991

2,535

5m: 2,535 / 1h: 0

缓存读取 0 0 14,884 22,810

响应

pong

pong

pong

pong

步骤 2: identity text

时间 17:39:17 ~ 17:39:27 (5.2s) 17:40:26 ~ 17:40:36 (6.2s) 17:39:21 ~ 17:39:26 (3.3s) 17:41:35 ~ 17:41:40 (2.8s)

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) 3,608 ms 4,583 ms 1,832 ms 1,752 ms

首文本(SSE delta) 3,610 ms 4,585 ms 1,834 ms 1,753 ms

总耗时 4,060 ms 5,193 ms 2,298 ms 1,821 ms

输入 3 3 3 26,803

输出 22 22 22 6

缓存创建

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

缓存读取 13,122 20,875 20,875 0

响应

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

Claude Sonnet 4.6

步骤 3: cutoff text

时间 17:41:07 ~ 17:41:12 (3.4s) 17:42:03 ~ 17:42:13 (6.9s) 17:41:15 ~ 17:41:20 (4.3s) 17:43:24 ~ 17:43:34 (8.8s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 2,176 ms 5,491 ms 3,138 ms 7,402 ms

首文本(SSE delta) 2,362 ms 5,652 ms 3,307 ms 7,596 ms

总耗时 2,459 ms 5,687 ms 3,390 ms 7,660 ms

输入 3 3 3 1,341

输出 8 8 8 2

缓存创建

5m: 0 / 1h: 38

5m: 0 / 1h: 57

5m: 0 / 1h: 38

2,549

5m: 2,549 / 1h: 0

缓存读取 13,311 21,064 21,064 22,939

响应

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 17:42:59 ~ 17:43:09 (6.1s) 17:43:59 ~ 17:44:10 (9.4s) 17:44:51 ~ 17:45:01 (5.0s) 17:45:46 ~ 17:45:51 (2.9s)

提示词

查看本步 Prompt

中文做个自我介绍：你是什么产品或助手？谁开发的？50 字内。

HTTP 头(TTFB) 4,312 ms 7,026 ms 3,126 ms 1,898 ms

首文本(SSE delta) 4,313 ms 7,034 ms 3,518 ms 1,899 ms

总耗时 5,059 ms 8,285 ms 4,004 ms 1,970 ms

输入 3 3 3 26,851

输出 67 61 62 16

缓存创建

5m: 0 / 1h: 43

缓存读取 13,349 21,121 21,102 0

输出速度 89.8 tok/s 48.8 tok/s 127.6 tok/s -

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助你完成写作、分析、编程、问答等各种任务，致力于提供安全、有益的智能对话服务。

我是 Claude，由 Anthropic 开发的 AI 助手。当前运行的模型版本是 claude-sonnet-4-6，知识截止日期为 2025 年 8 月。

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

我是 Claude Sonnet 4.6（claude-sonnet-4-6）。

步骤 5: knowledge_recall text

时间 17:44:57 ~ 17:45:12 (14.7s) 17:47:35 ~ 17:47:45 (6.6s) 17:46:04 ~ 17:46:09 (4.3s) 17:48:16 ~ 17:48:26 (5.4s)

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) 3,231 ms 3,332 ms 1,490 ms 2,159 ms

首文本(SSE delta) 4,445 ms 4,413 ms 1,492 ms 2,198 ms

总耗时 13,703 ms 5,387 ms 3,328 ms 4,391 ms

输入 3 3 3 580

输出 844 205 204 158

缓存创建

5,070

5m: 0 / 1h: 5,070

8,813

5m: 0 / 1h: 8,813

6,664

5m: 0 / 1h: 6,664

1,103

5m: 1,103 / 1h: 0

缓存读取 11,447 0 2,131 9,924

输出速度 91.2 tok/s burst 111.1 tok/s 72.0 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

{"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 17:48:46 ~ 17:49:16 (28.8s) 17:50:02 ~ 17:50:22 (18.9s) 17:49:42 ~ 17:50:02 (17.7s) 17:49:29 ~ 17:49:44 (12.2s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=143966994775126565642634458064734547751615595784484144899196281257593717

HTTP 头(TTFB) 7,270 ms 1,562 ms 1,256 ms 1,652 ms

首文本(SSE delta) 8,221 ms 1,566 ms 1,261 ms 1,653 ms

总耗时 27,823 ms 17,776 ms 16,777 ms 11,182 ms

输入 3 164 164 12

输出 1,938 1,423 1,606 198

缓存创建

452

5m: 0 / 1h: 452

5m: 0 / 1h: 0

5m: 22 / 1h: 0

缓存读取 13,202 0 0 198

输出速度 98.9 tok/s 87.8 tok/s 103.5 tok/s 20.8 tok/s

响应

I'll count character by character, grouping in tens:

```
s = 143966994775126565642634458064734547751615595784484144899196281257593717
```

**Segmented:**
```
1439669947  (1–10)
7512656564  (11–20)
26

## Counting Digits in s

**s** = `143966994775126565642634458064734547751615595784484144899196281257593717`

### Step 1 — Total Length

Splitting into groups of 4:
`1439 | 6699 | 4775 | 1265 | 6564 |

I'll work through the string systematically.

**s = `143966994775126565642634458064734547751615595784484144899196281257593717`**

**Length — indexing each character:**

```
 1: 1   2: 4   3: 3   4: 9

```json
{"n":72,"c":[1,9,8]}
```

总计

总输入(含缓存) 83,363 93,216 93,160 119,001

总输出 2,884 1,724 1,907 381

总缓存创建

18,914

5m: 0 / 1h: 18,914

29,977

5m: 0 / 1h: 29,977

12,925

5m: 0 / 1h: 12,925

6,209

5m: 6,209 / 1h: 0

总缓存读取 64,431 63,060 80,056 55,871

输出速度 96.2 tok/s 85.0 tok/s 104.9 tok/s 30.4 tok/s

总执行时间 1m 4.3s 55.5s 38.6s 38.8s

总墙钟时间

11m 33.0s

含步骤间等待

12m 41.0s

含步骤间等待

12m 14.9s

含步骤间等待

11m 58.9s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1336

Anthropic: Claude Sonnet 4.6

$0.1577

Anthropic: Claude Sonnet 4.6

$0.1016

Anthropic: Claude Sonnet 4.6

$0.2078

Anthropic: Claude Sonnet 4.6

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 one-api (100%) one-api (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 - cloudflare

ID 格式 msg_01NGqzZPbEKWM48wCRhw2pSo msg_01Fk3szcMYa6zFM1NvY9dK6c msg_01C4879jqqGTWhGX9Ta3NFkA msg_1775b26ea4d349c390ef4def09783c49

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-NL	via: 1.1 Caddy cf-ray: a14cbc9bfe60e7b2-FRA server: cloudflare x-new-api-version: v0.12.10 x-oneapi-request-id: 202607020948474410136278268d9d6bUcjNXFX	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: false usage_has_cache_fields: true	hdr:x-new-api-versionhdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-rayhdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14cbc9bfe60e7b2-FRA
O-Max	cf-ray: a14cbe780bc6a017-AMS server: cloudflare x-new-api-version: 0.12.11 x-oneapi-request-id: 202607020950037650670968268d9d6gGGEAzeB	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-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14cbe780bc6a017-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: 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_011Ccd2F4hMSLpMC5xwzc6Hb
M-Kiro/Antigravity	cf-ray: a14cbda91950971a-AMS server: cloudflare x-oneapi-request-id: 202607020949305191362308268d9d6iqhQJTGP	id_format: msg_prefix extra_fields: ["container"] has_service_tier: true has_inference_geo: false 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.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14cbda91950971a-AMS

协议指纹评分（机器自动）

总分 / 100 94 92 100 51

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 9.6 /10 ×20 10.0 /10 ×20 1.5 /10 ×20

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

延迟基线 9.8 /10 ×15 5.4 /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 2.3 /10 ×13

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

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

身份(结构) 10.0 /10 ×7 10.0 /10 ×7 10.0 /10 ×7 0.0 /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 4.0 /10 ×6

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

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

流式投递 10.0 /10 ×5 3.2 /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 0.0 /10 ×2

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

综合结论

与基线相似度 94 v3.26.0 92 v3.26.0 基准（参考） 51 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 流式投递 buffer_dump_match
6 自由身份 identity_free_clean
7 停止原因 stop_reason_present
8 系统提示纯净 system_prompt_clean
9 Service Tier service_tier_present
10 缓存命中比 cache_hit_ratio_match
11 缓存 TTL 一致 cache_ttl_consistency
12 推理地区 inference_geo_present
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 9659 ms 8121 ms 5458 ms 5162 ms

输出速度 96.2 tok/s 85.0 tok/s 104.9 tok/s 30.4 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 ⓘ

$0.1336

$0.1577

$0.1016

$0.2078

缓存读取占比 ⓘ 77% 68% 86% 47%

揭晓

通道

服务商 sucui ↗

通道O-Max-NL

服务商 Aiide ↗

通道O-Max

官方基准

服务商 luminai ↗

通道M-Kiro/Antigravity

返回列表