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

DBAI / M-
```
评测失败
```

步骤 / 指标

官方基准

SAIAi

O-Max

SudoCode

M-Max

FoxCode

R-ultra

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 12:24:05 ~ 12:24:10 (3.6s) 12:23:59 ~ 12:24:04 (4.6s) 12:24:03 ~ 12:24:18 (10.2s) 12:24:01 ~ 12:24:06 (4.7s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,107 ms 3,409 ms 8,699 ms 2,735 ms

首文本(SSE delta) 2,108 ms 3,411 ms 8,701 ms 2,736 ms

总耗时 2,601 ms 3,557 ms 8,958 ms 3,680 ms

输入 3 3 3 43

输出 5 5 5 14

缓存创建

5m: 0 / 1h: 0

32,318

5m: 0 / 1h: 32,318

32,318

5m: 0 / 1h: 32,318

32,972

5m: 32,972 / 1h: 0

缓存读取 32,318 0 0 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 12:27:12 ~ 12:27:22 (6.8s) 12:25:26 ~ 12:25:31 (3.1s) 12:26:29 ~ 12:26:44 (10.6s) 12:26:23 ~ 12:26:33 (4.9s)

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。请按下列 3 行格式精确填空作答（label 不变，值小写）：
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) 5,313 ms 1,780 ms 9,214 ms 3,561 ms

首文本(SSE delta) 5,314 ms 1,781 ms 9,215 ms 3,577 ms

总耗时 5,839 ms 2,123 ms 9,605 ms 4,128 ms

输入 3 3 3 43

输出 22 22 22 123

缓存创建

5m: 0 / 1h: 0

182

5m: 0 / 1h: 182

32,500

5m: 0 / 1h: 32,500

33,484

5m: 33,484 / 1h: 0

缓存读取 32,500 32,318 0 0

输出速度 - - - burst

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-4-6

步骤 3: cutoff text

时间 12:30:05 ~ 12:30:10 (4.7s) 12:27:14 ~ 12:27:19 (3.6s) 12:30:37 ~ 12:30:47 (9.0s) 12:28:51 ~ 12:29:01 (5.9s)

提示词

查看本步 Prompt

你掌握的世界知识截止于何时？YYYY-MM 格式作答，不解释。

HTTP 头(TTFB) 3,066 ms 2,314 ms 7,621 ms 4,061 ms

首文本(SSE delta) 3,284 ms 2,316 ms 7,623 ms 4,062 ms

总耗时 3,410 ms 2,529 ms 7,901 ms 4,785 ms

输入 3 3 3 89

输出 8 8 8 33

缓存创建

5m: 0 / 1h: 55

5m: 0 / 1h: 57

239

5m: 0 / 1h: 239

411

5m: 411 / 1h: 0

缓存读取 32,500 32,500 32,318 33,484

输出速度 - - - 45.6 tok/s

响应

2025-08

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 12:32:41 ~ 12:32:51 (5.6s) 12:30:35 ~ 12:30:45 (5.0s) 12:34:45 ~ 12:34:55 (9.6s) 12:32:09 ~ 12:32:19 (5.7s)

提示词

查看本步 Prompt

中文回答：你叫什么、属于哪个产品、谁开发？50 字内。

HTTP 头(TTFB) 3,881 ms 2,775 ms 8,019 ms 3,735 ms

首文本(SSE delta) 3,883 ms 2,783 ms 8,021 ms 3,736 ms

总耗时 4,600 ms 3,957 ms 8,343 ms 4,707 ms

输入 3 3 3 67

输出 46 57 26 70

缓存创建

5m: 0 / 1h: 41

5m: 0 / 1h: 43

32,598

5m: 0 / 1h: 32,598

34,278

5m: 34,278 / 1h: 0

缓存读取 32,555 32,557 0 0

输出速度 64.2 tok/s 48.6 tok/s - 72.1 tok/s

响应

我叫 Claude，是由 Anthropic 开发的 AI 助手。当前使用的模型版本为 claude-sonnet-4-6。

我是 Claude，由 Anthropic 开发的 AI 助手。当前版本为 Claude Sonnet 4.6，擅长对话、编程、分析和写作等任务。

我叫 Claude，是由 Anthropic 开发的 AI 助手。

我叫 Kiro，基于 Anthropic 开发的 Claude 模型（claude-sonnet-4-6）构建，是AWS 的 AI 编程助手产品。

步骤 5: knowledge_recall text

时间 12:34:34 ~ 12:34:44 (5.3s) 12:33:27 ~ 12:33:37 (5.1s) 12:38:22 ~ 12:38:32 (6.8s) 12:34:14 ~ 12:34:39 (23.4s)

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) 1,868 ms 1,687 ms 3,197 ms 2,298 ms

首文本(SSE delta) 1,870 ms 1,688 ms 3,199 ms 2,472 ms

总耗时 3,610 ms 4,162 ms 5,729 ms 22,048 ms

输入 3 3 3 21

输出 196 196 196 1,817

缓存创建

2,313

5m: 0 / 1h: 2,313

8,046

5m: 0 / 1h: 8,046

8,013

5m: 0 / 1h: 8,013

9,715

5m: 9,715 / 1h: 0

缓存读取 5,718 0 0 0

输出速度 112.6 tok/s 79.2 tok/s 77.5 tok/s 92.8 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 12:36:31 ~ 12:36:56 (20.5s) 12:35:02 ~ 12:35:22 (16.8s) 12:40:16 ~ 12:40:41 (20.4s) 12:36:28 ~ 12:36:58 (25.7s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=971530025736103762113245917305747216947895608738674748781042077593163902

HTTP 头(TTFB) 1,221 ms 1,540 ms 1,994 ms 2,512 ms

首文本(SSE delta) 1,606 ms 1,541 ms 2,018 ms 2,516 ms

总耗时 19,370 ms 15,768 ms 19,359 ms 24,971 ms

输入 164 164 164 64

输出 1,548 1,139 1,459 3,109

缓存创建

5m: 0 / 1h: 0

144

5m: 144 / 1h: 0

缓存读取 0 0 0 0

输出速度 87.1 tok/s 80.1 tok/s 84.1 tok/s 138.5 tok/s

响应

```
s = 971530025736103762113245917305747216947895608738674748781042077593163902
```

**Digit-by-digit indexing (groups of 10):**

| # | Digits |
|---|--------|
| 1–10 | 9 7 1 5 3 **0 0** 2 5 **7** |

## Counting digits in:
`971530025736103762113245917305747216947895608738674748781042077593163902`

### Length
Splitting into groups of 10:
```
9715300257 | 3610376211 | 3245917305 | 7472169478 |
95608

## Counting digits in s

**s:** `971530025736103762113245917305747216947895608738674748781042077593163902`

### Length
Grouping in tens:
```
9715300257 | 3610376211 | 3245917305 | 7472169478 | 9560873

{"n":72,"c":[8,13,6]}

总计

总输入(含缓存) 138,179 138,200 138,165 144,815

总输出 1,825 1,427 1,716 5,166

总缓存创建

2,409

5m: 0 / 1h: 2,409

40,646

5m: 0 / 1h: 40,646

105,668

5m: 0 / 1h: 105,668

111,004

5m: 111,004 / 1h: 0

总缓存读取 135,591 97,375 32,318 33,484

输出速度 88.5 tok/s 77.9 tok/s 83.3 tok/s 115.0 tok/s

总执行时间 46.5s 38.2s 1m 6.6s 1m 10.4s

总墙钟时间

12m 50.4s

含步骤间等待

11m 22.8s

含步骤间等待

16m 37.2s

含步骤间等待

12m 56.5s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0776

Anthropic: Claude Sonnet 4.6

$0.2036

Anthropic: Claude Sonnet 4.6

$0.4322

Anthropic: Claude Sonnet 4.6

$0.3487

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%) direct (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_016B8c8scRKr3xRPFsd4RBAy msg_01H46jgoR5jXJaot6gddTmnM msg_0142hWfy4vSFf4e3TwbSaScA msg_01W0vfOz2QtwckC7scy40kUX

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_011Cc8DE9yzXeTKYKjWF7LmJ
O-Max	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_011Cc8D7ezZ5oYdJwBPQrWue
M-Max	server: nginx x-oneapi-request-id: 20260617124017149609687L37pPEtS	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	—
R-ultra	cf-ray: a0cf5a8088427343-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606170436296955835248268d9d6tVC8WiG1	id_format: msg_prefix 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_fieldsupstream:anthropic-direct-from-body	a0cf5a8088427343-NRT

协议指纹评分（机器自动）

总分 / 100 100 97 81 72

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 7.3 /10 ×20 0.0 /10 ×20 0.0 /10 ×20

缓存 TTL 一致 10.0 /10 ×15 10.0 /10 ×15 10.0 /10 ×15 4.0 /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 5.3 /10 ×13 5.3 /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 5.0 /10 ×8

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

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 7.8 /10 ×5 10.0 /10 ×5

Req-ID 透传 10.0 /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

与基线相似度基准（参考） 97 v3.21.0 81 v3.21.0 72 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 6572 ms 5349 ms 9983 ms 10720 ms

输出速度 88.5 tok/s 77.9 tok/s 83.3 tok/s 115.0 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 ⓘ

$0.0776

$0.2036

$0.4322

$0.3487

缓存读取占比 ⓘ 98% 70% 23% 23%

揭晓

通道

官方基准

服务商 SAIAi ↗

通道O-Max

服务商 SudoCode ↗

通道M-Max

服务商 FoxCode ↗

通道R-ultra

返回列表