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

2026/06/22 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

InfAI / R-Kiro
```
评测失败
```

步骤 / 指标

官方基准

ClaudeAPI

O-Bdrk

HongMaCC

u-cc

OmniaKey

O-Max

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 v2.1.160 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 07:21:48 ~ 07:21:53 (4.7s) 07:21:42 ~ 07:21:52 (6.8s) 07:21:46 ~ 07:21:56 (7.2s) 07:21:44 ~ 07:21:54 (6.8s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 3,232 ms 4,771 ms 5,557 ms 5,187 ms

首文本(SSE delta) 3,234 ms 4,790 ms 5,559 ms 5,189 ms

总耗时 3,336 ms 5,001 ms 5,559 ms 5,328 ms

输入 3 3 3 3

输出 5 5 5 5

缓存创建

32,318

5m: 0 / 1h: 32,318

31,628

5m: 31,628 / 1h: 0

26,613

5m: 0 / 1h: 26,613

32,318

5m: 0 / 1h: 32,318

缓存读取 0 0 0 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 07:23:59 ~ 07:24:04 (3.8s) 07:23:50 ~ 07:24:00 (5.8s) 07:24:26 ~ 07:24:36 (7.6s) 07:25:51 ~ 07:26:01 (5.4s)

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) 2,088 ms 4,062 ms 6,375 ms 3,869 ms

首文本(SSE delta) 2,089 ms 4,063 ms 6,376 ms 3,870 ms

总耗时 2,630 ms 4,439 ms 6,376 ms 4,345 ms

输入 3 3 3 3

输出 22 22 22 22

缓存创建

189

5m: 0 / 1h: 189

189

5m: 189 / 1h: 0

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

缓存读取 32,318 31,628 26,613 32,318

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-4-5

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

步骤 3: cutoff text

时间 07:26:02 ~ 07:26:07 (4.6s) 07:27:54 ~ 07:28:09 (10.8s) 07:26:11 ~ 07:26:21 (5.2s) 07:29:44 ~ 07:29:54 (5.4s)

提示词

查看本步 Prompt

Knowledge cutoff date? Reply with YYYY-MM only.

HTTP 头(TTFB) 3,212 ms 8,793 ms 4,073 ms 3,763 ms

首文本(SSE delta) 3,336 ms 9,236 ms 4,074 ms 3,953 ms

总耗时 3,493 ms 9,565 ms 4,074 ms 4,080 ms

输入 3 3 3 3

输出 8 8 8 8

缓存创建

5m: 0 / 1h: 38

5m: 57 / 1h: 0

5m: 0 / 1h: 57

5m: 0 / 1h: 38

缓存读取 32,507 31,817 26,802 32,507

响应

2025-08

2025-08

2025-04

2025-08

步骤 4: identity_free text

时间 07:27:21 ~ 07:27:26 (3.8s) 07:31:23 ~ 07:31:33 (5.7s) 07:28:26 ~ 07:28:36 (6.1s) 07:31:21 ~ 07:31:31 (5.9s)

提示词

查看本步 Prompt

请用中文介绍你自己是什么产品或助手，由谁开发，50 字以内。

HTTP 头(TTFB) 1,837 ms 3,476 ms 4,982 ms 3,825 ms

首文本(SSE delta) 2,225 ms 3,529 ms 4,984 ms 3,826 ms

总耗时 2,857 ms 4,177 ms 4,984 ms 4,750 ms

输入 3 3 3 3

输出 62 53 69 55

缓存创建

5m: 0 / 1h: 43

5m: 41 / 1h: 0

28,673

5m: 0 / 1h: 28,673

5m: 0 / 1h: 35

缓存读取 32,545 31,874 0 32,545

输出速度 98.1 tok/s 81.8 tok/s burst 59.5 tok/s

响应

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。

我叫 Claude，是由 Anthropic 开发的 AI 助手，当前模型版本为 claude-sonnet-4-6，属于 Claude 产品系列。

我是 Claude，由 Anthropic 开发的 AI 助手。我是 Claude Code，专为软件开发场景设计的命令行工具，帮助开发者完成编程、调试、代码审查等任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我基于大型语言模型构建，旨在提供安全、有益的对话和任务协助。

步骤 5: knowledge_recall text

时间 07:29:47 ~ 07:29:52 (4.7s) 07:34:03 ~ 07:34:13 (5.2s) 07:31:27 ~ 07:31:42 (10.8s) 07:35:21 ~ 07:35:31 (6.1s)

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) 1,611 ms 2,108 ms 9,387 ms 3,030 ms

首文本(SSE delta) 1,613 ms 2,161 ms 9,388 ms 3,031 ms

总耗时 3,501 ms 3,958 ms 9,388 ms 4,972 ms

输入 3 3 3 3

输出 204 198 524 204

缓存创建

8,039

5m: 0 / 1h: 8,039

8,073

5m: 8,073 / 1h: 0

19,284

5m: 0 / 1h: 19,284

8,024

5m: 0 / 1h: 8,024

缓存读取 0 0 0 0

输出速度 108.1 tok/s 110.2 tok/s burst 105.1 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 07:30:59 ~ 07:31:19 (18.0s) 07:38:02 ~ 07:40:37 (154.0s) 07:34:41 ~ 07:35:06 (24.0s) 07:39:21 ~ 07:39:41 (15.4s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=702699637956051588417260555106234056689892531737011691668327671828119314

HTTP 头(TTFB) 1,057 ms 142,237 ms 22,831 ms 5,937 ms

首文本(SSE delta) 1,059 ms 142,293 ms 22,832 ms 5,984 ms

总耗时 17,014 ms 152,720 ms 22,832 ms 13,854 ms

输入 164 191 3 164

输出 1,601 1,106 1,962 756

缓存创建

5m: 0 / 1h: 0

452

5m: 0 / 1h: 452

5m: 0 / 1h: 0

缓存读取 0 0 16,620 0

输出速度 100.3 tok/s 106.1 tok/s burst 96.1 tok/s

响应

## Counting digits in s

**s = 702699637956051588417260555106234056689892531737011691668327671828119314**

### Length (n)

Splitting into groups of 10:
```
7026996379 | 5605158841 | 7260555106 | 23405

{"n":72,"c":[6,7,7]}

I'll work through the string methodically, indexing each character:

**s** = `702699637956051588417260555106234056689892531737011691668327671828119314`

---

**Step 1 – Total length (n)**

Breaking in

```json
{"n":72,"c":[6,7,7]}
```

**Verification:**

String: `702699637956051588417260555106234056689892531737011691668327671828119314`

- **Length (n):** 72
- **0s:** positions 2,13,24,29,34,49 → **6

总计

总输入(含缓存) 138,176 135,513 145,321 138,153

总输出 1,902 1,392 2,590 1,050

总缓存创建

40,627

5m: 0 / 1h: 40,627

39,988

5m: 39,988 / 1h: 0

75,268

5m: 0 / 1h: 75,268

40,604

5m: 0 / 1h: 40,604

总缓存读取 97,370 95,319 70,035 97,370

输出速度 101.1 tok/s 105.4 tok/s burst 94.6 tok/s

总执行时间 39.5s 3m 8.3s 1m 0.8s 45.1s

总墙钟时间

9m 30.9s

含步骤间等待

18m 55.3s

含步骤间等待

13m 20.1s

含步骤间等待

17m 57.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.2106

Anthropic: Claude Sonnet 4.6

$0.1438

Anthropic: Claude Sonnet 4.6

$0.3422

Anthropic: Claude Sonnet 4.6

$0.1978

Anthropic: Claude Sonnet 4.6

通道指纹

来源步骤 digit_count digit_count digit_count digit_count

平台 direct (100%) one-api (100%) one-api (100%) one-api (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_01MipUsLZ7PjbzUMa2HxJcsx msg_01SNnetAH9qAi7SQUZvPBt4w msg_01J6482eXBT7sf8uXESvHQA5 msg_019F2Shd2DTVfn7XPkNieNF7

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
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_011CcHGyQafNNB34u7uMBNJ4
O-Bdrk	x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606212338037140172648268d9d6Vi7tSnmq	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	—
u-cc	server: nginx x-new-api-version: 1.0.14 x-oneapi-request-id: 202606212334424790941708268d9d6fsavNhE2	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	—
O-Max	cf-ray: a0f6da2c2907bdf3-AMS server: cloudflare x-oneapi-request-id: 202606212339233839134428268d9d6dl7cFuZH	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	a0f6da2c2907bdf3-AMS

协议指纹评分（机器自动）

总分 / 100 100 86 73 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 3.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 0.0 /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 7.7 /10 ×13 10.0 /10 ×13

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

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

知识截止 10.0 /10 ×7 10.0 /10 ×7 5.0 /10 ×7 10.0 /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 10.0 /10 ×6

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

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

流式投递 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5 10.0 /10 ×5

Req-ID 透传 10.0 /10 ×4 0.0 /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

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

综合结论

与基线相似度基准（参考） 86 v3.24.0 73 v3.24.0 98 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 流式投递 buffer_dump_match
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 5472 ms 29977 ms 8869 ms 6222 ms

输出速度 101.1 tok/s 105.4 tok/s burst 94.6 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 ⓘ

$0.2106

$0.1438

$0.3422

$0.1978

缓存读取占比 ⓘ 70% 70% 48% 70%

揭晓

通道

官方基准

服务商 ClaudeAPI ↗

通道O-Bdrk

服务商 HongMaCC ↗

通道u-cc

服务商 OmniaKey ↗

通道O-Max

返回列表