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

2026/06/17 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

⚠ 1 通道不可测 · 已从主对比表移除点击展开原因

ZeroCode / cc
```
评测失败
```

步骤 / 指标

SudoCode

M-Max

官方基准

0-0

O-Max

DawAPI

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 16:17:20 ~ 16:17:26 (4.4s) 16:17:22 ~ 16:17:28 (4.7s) 16:17:18 ~ 16:21:49 (266.1s) 16:17:16 ~ 16:17:27 (5.2s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,883 ms 3,307 ms 125,488 ms 3,484 ms

首文本(SSE delta) 2,889 ms 3,309 ms - 3,485 ms

总耗时 3,019 ms 3,584 ms 125,489 ms 3,650 ms

输入 3 3 3 3

输出 5 5 5 5

缓存创建

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

32,310

5m: 32,310 / 1h: 0

27,260

5m: 27,260 / 1h: 0

缓存读取 - 0 0 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 16:19:21 ~ 16:19:26 (4.3s) 16:18:28 ~ 16:18:33 (3.7s) 16:24:01 ~ 16:24:11 (6.5s) 16:18:44 ~ 16:18:54 (5.0s)

提示词

查看本步 Prompt

以下是对你自身的 metadata 探针。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。按以下 3 行回答自己的身份，每行一个 label，不要寒暄不要解释：
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) 2,395 ms 2,172 ms 4,683 ms 3,513 ms

首文本(SSE delta) 2,416 ms 2,174 ms 4,684 ms 3,514 ms

总耗时 3,011 ms 2,529 ms 5,089 ms 3,836 ms

输入 3 3 3 3

输出 22 22 22 22

缓存创建

202

5m: 0 / 1h: 202

202

5m: 0 / 1h: 202

202

5m: 202 / 1h: 0

202

5m: 202 / 1h: 0

缓存读取 32,318 32,318 32,310 27,260

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

vendor: anthropic
brand: claude
model: claude-sonnet-4-5

步骤 3: cutoff text

时间 16:22:47 ~ 16:22:57 (5.1s) 16:21:43 ~ 16:21:48 (3.2s) 16:26:03 ~ 16:26:13 (7.5s) 16:20:18 ~ 16:20:23 (4.4s)

提示词

查看本步 Prompt

请直接回答你的知识截止日期，格式 YYYY-MM，不要附加任何说明。

HTTP 头(TTFB) 3,814 ms 1,954 ms 5,318 ms 3,046 ms

首文本(SSE delta) 3,815 ms 2,094 ms 5,320 ms 3,047 ms

总耗时 4,082 ms 2,164 ms 5,939 ms 3,674 ms

输入 3 3 3 3

输出 8 8 8 8

缓存创建

32,577

5m: 0 / 1h: 32,577

5m: 0 / 1h: 38

5m: 57 / 1h: 0

5m: 55 / 1h: 0

缓存读取 - 32,520 32,512 27,462

响应

2025-08

2025-08

2025-08

2024-04

步骤 4: identity_free text

时间 16:24:09 ~ 16:24:14 (4.6s) 16:25:27 ~ 16:25:32 (4.7s) 16:28:28 ~ 16:28:38 (8.8s) 16:23:56 ~ 16:24:06 (8.8s)

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) 2,697 ms 2,662 ms 6,394 ms 6,935 ms

首文本(SSE delta) 2,706 ms 3,064 ms 6,395 ms 6,936 ms

总耗时 3,453 ms 3,709 ms 7,100 ms 8,068 ms

输入 3 3 3 3

输出 58 58 52 65

缓存创建

100

5m: 0 / 1h: 100

5m: 0 / 1h: 35

5m: 41 / 1h: 0

5m: 43 / 1h: 0

缓存读取 32,520 32,558 32,569 27,517

输出速度 77.6 tok/s 89.9 tok/s 73.8 tok/s 57.4 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。我基于 claude-sonnet-4-6 模型，擅长编程、分析、写作和问答等任务。

我是 Claude，由 Anthropic 开发的 AI 助手。我基于大型语言模型构建，旨在提供有帮助、无害且诚实的对话服务。

我叫 Claude，是 Anthropic 开发的 AI 助手。当前版本为 Claude Sonnet 4.6，属于 Claude 4 系列产品。

我是 Claude，由 Anthropic 开发的 AI 助手。我能帮助编写代码、分析问题、回答问题等。目前作为 Claude Code 的底层模型为你提供服务。

步骤 5: knowledge_recall text

时间 16:27:48 ~ 16:27:58 (5.4s) 16:29:32 ~ 16:29:43 (5.3s) 16:32:36 ~ 16:32:46 (7.9s) 16:27:50 ~ 16:28:01 (8.0s)

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) 1,825 ms 2,345 ms 3,580 ms 4,203 ms

首文本(SSE delta) 1,826 ms 2,346 ms 3,582 ms 4,204 ms

总耗时 4,290 ms 4,289 ms 6,201 ms 6,938 ms

输入 3 3 3 3

输出 205 204 205 198

缓存创建

8,067

5m: 0 / 1h: 8,067

8,040

5m: 0 / 1h: 8,040

8,051

5m: 8,051 / 1h: 0

3,217

5m: 3,217 / 1h: 0

缓存读取 - 0 0 0

输出速度 83.2 tok/s 105.0 tok/s 78.3 tok/s 72.4 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 16:29:27 ~ 16:29:47 (16.7s) 16:31:26 ~ 16:31:51 (24.5s) 16:35:30 ~ 16:37:25 (114.9s) 16:30:46 ~ 16:31:11 (24.3s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=665734713495107925830600813394803627498725375384036793284416200710798798

HTTP 头(TTFB) 1,609 ms 1,626 ms 81,119 ms 2,773 ms

首文本(SSE delta) 1,630 ms 1,629 ms - 20,079 ms

总耗时 15,499 ms 23,623 ms 81,120 ms 23,440 ms

输入 164 164 162 1,001

输出 1,022 1,951 892 2,137

缓存创建

5m: 0 / 1h: 0

缓存读取 - 0 0 0

输出速度 73.7 tok/s 88.7 tok/s burst burst

响应

```json
{"n":72,"c":[9,10,7]}
```

## Counting digits in s

**s = `665734713495107925830600813394803627498725375384036793284416200710798798`**

### Step 1 — Length

Splitting into groups of 10:
```
6657347134  → 10
9510792583  → 10
060

I need to count the digits in the string carefully.

s = 665734713495107925830600813394803627498725375384036793284416200710798798

First, let me find the length, then count 0s, 7s, and 9s.

**String:*

Breaking `s` into groups of 10 for accuracy:

```
6657347134  9510792583  0600813394  8036274987
2537538403  6793284416  2007107987  98
```

**Length:** 8×10 + 2 = **72**

**Count 0s** (scanning each

总计

总输入(含缓存) 138,281 138,208 138,229 114,032

总输出 1,320 2,248 1,184 2,435

总缓存创建

73,264

5m: 0 / 1h: 73,264

40,633

5m: 0 / 1h: 40,633

40,661

5m: 40,661 / 1h: 0

30,777

5m: 30,777 / 1h: 0

总缓存读取 64,838 97,396 97,391 82,239

输出速度 75.2 tok/s 90.0 tok/s 77.3 tok/s 68.0 tok/s

总执行时间 40.5s 46.0s 6m 51.7s 55.8s

总墙钟时间

12m 26.4s

含步骤间等待

14m 28.3s

含步骤间等待

20m 6.8s

含步骤间等待

13m 54.3s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3145

Anthropic: Claude Sonnet 4.6

$0.2158

Anthropic: Claude Sonnet 4.6

$0.1428

Anthropic: Claude Sonnet 4.6

$0.1364

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

上游 - Anthropic 直连 (100%) Anthropic 直连 (100%) Anthropic 直连 (100%)

CDN - - - -

ID 格式 msg_01SBVX9Q21ZxaHyYytU5c2xW msg_011QLbmvLpaywh2L8pvCKvUU msg_01NErufvvfK53BHuGo5hesyC msg_013BWKuVP8Zk8rTEHjvZcH75

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
M-Max	server: nginx x-oneapi-request-id: 20260617162928591683339Q68Pv6Ux	id_format: msg_prefix has_service_tier: false has_inference_geo: false model_has_date_suffix: false usage_has_cache_fields: false	hdr:x-oneapi-request-idbody:id:msg_prefix	f136305d864c2ec6ca2d3c6182485add
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_011Cc8X9APkaZ63vR8JBBTFp
O-Max	via: 1.1 Caddy server: uvicorn, uvicorn, nginx/1.24.0 (Ubuntu)	id_format: msg_prefix via_chain: ["1.1 Caddy"] extra_fields: ["context_management","stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: false usage_has_cache_fields: true	hdr:viabody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	908b7f9693e329e9f8d4080c2b8e7299
cc	server: cdn x-new-api-version: v1.0.0-rc.11 x-oneapi-request-id: 202606170830471957220668268d9d6zZ72iuXe	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	—

协议指纹评分（机器自动）

总分 / 100 75 98 86 71

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 5.5 /10 ×20 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 0.0 /10 ×15 10.0 /10 ×15 0.0 /10 ×15 0.0 /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 7.7 /10 ×13 10.0 /10 ×13 10.0 /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 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 6.7 /10 ×7

知识截止 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 0.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6 10.0 /10 ×6

inference_geo 0.0 /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 2.3 /10 ×5 10.0 /10 ×5

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

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 75 v3.21.0 基准（参考） 86 v3.21.0 71 v3.21.0

雷达图例

共 20 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 自由身份 identity_free_clean
5 停止原因 stop_reason_present
6 系统提示纯净 system_prompt_clean
7 Service Tier service_tier_present
8 缓存命中比 cache_hit_ratio_match
9 缓存 TTL 一致 cache_ttl_consistency
10 推理地区 inference_geo_present
11 缓存连续 cache_continuity_intra
12 知识召回 knowledge_recall_match
13 延迟基线 latency_baseline_match
14 原生 msg-ID anthropic_msg_id_format
15 Sliding 缓存 cache_sliding_correctness
16 结构化身份 identity_structured_match
17 信封自报一致 envelope_self_report_match
18 中文遵循 instruction_following_lang
19 档位思考量 tier_thinking_volume_match
20 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 5559 ms 6650 ms 38490 ms 8268 ms

输出速度 75.2 tok/s 90.0 tok/s 77.3 tok/s 68.0 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 ⓘ

$0.3145

$0.2158

$0.1428

$0.1364

缓存读取占比 ⓘ 47% 70% 70% 72%

揭晓

通道

服务商 SudoCode ↗

通道M-Max

官方基准

服务商 0-0 ↗

通道O-Max

服务商 DawAPI ↗

通道cc

返回列表