模型对比 | RPDiag

测试说明点击展开 5 步 quick-probe-v1 详解

前 4 步同一会话（交互 TUI 多轮），第 5、6 步各自独立新会话，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 1-3 步的延迟在交互进程内发生，让缓存前缀在墙钟里老化）。第 3 步（cutoff）累计跨度 > 5 分钟保住 sliding cache 检测；总跨度 > 6 分钟。第 5 步（知识召回）与第 6 步（档位）都是独立新会话（第 6 步不复用上下文，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

2 任务失败执行不完整

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

不可测：未取得可评分响应（如 403 / 上游错误）

步骤 / 指标

SSSAiCode

O-Max-L2

LinkAPI

O-Max

官方基准

sucui

O-Max-NL

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 23:47:14 ~ 23:54:35 (437.9s) 23:46:03 ~ 23:54:03 (476.4s) 23:49:13 ~ 23:58:14 (537.6s) 23:47:47 ~ 23:57:08 (556.3s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 6,853 ms 2,811 ms 2,777 ms 2,775 ms

首文本(SSE delta) 6,939 ms 2,812 ms 2,778 ms 2,776 ms

总耗时 7,032 ms 3,015 ms 2,884 ms 3,036 ms

输入 3 3 3 3

输出 5 5 5 5

缓存创建

21,899

5m: 0 / 1h: 21,899

22,307

5m: 0 / 1h: 22,307

5m: 0 / 1h: 0

14,507

5m: 0 / 1h: 14,507

缓存读取 0 0 22,263 0

响应

pong

pong

pong

pong

步骤 2: identity text

时间 23:47:14 ~ 23:54:35 (437.9s) 23:46:03 ~ 23:54:03 (476.4s) 23:49:13 ~ 23:58:14 (537.6s) 23:47:47 ~ 23:57:08 (556.3s)

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) 2,159 ms 3,223 ms 2,361 ms 2,615 ms

首文本(SSE delta) 2,468 ms 3,225 ms 2,362 ms 2,616 ms

总耗时 2,593 ms 3,760 ms 2,777 ms 3,142 ms

输入 3 3 3 3

输出 22 22 22 22

缓存创建

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

189

5m: 0 / 1h: 189

缓存读取 21,899 22,307 22,263 14,507

响应

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

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 23:47:14 ~ 23:54:35 (437.9s) 23:46:03 ~ 23:54:03 (476.4s) 23:49:13 ~ 23:58:14 (537.6s) 23:47:47 ~ 23:57:08 (556.3s)

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) 2,521 ms 1,837 ms 4,500 ms 3,295 ms

首文本(SSE delta) 2,760 ms 1,966 ms 4,652 ms 3,486 ms

总耗时 2,962 ms 2,010 ms 4,710 ms 3,574 ms

输入 3 3 3 3

输出 8 8 8 8

缓存创建

5m: 0 / 1h: 55

5m: 0 / 1h: 57

缓存读取 22,088 22,496 22,452 14,696

响应

2025-04

2025-08

2025-08

2025-08

步骤 4: identity_free text

时间 23:47:14 ~ 23:54:35 (437.9s) 23:46:03 ~ 23:54:03 (476.4s) 23:49:13 ~ 23:58:14 (537.6s) 23:47:47 ~ 23:57:08 (556.3s)

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) 2,474 ms 2,976 ms 2,020 ms 2,787 ms

首文本(SSE delta) 2,650 ms 3,378 ms 2,021 ms 3,251 ms

总耗时 2,707 ms 3,965 ms 2,724 ms 3,975 ms

输入 3 3 3 3

输出 24 60 44 74

缓存创建

5m: 0 / 1h: 35

5m: 0 / 1h: 39

5m: 0 / 1h: 43

缓存读取 22,143 22,551 22,509 14,753

输出速度 - 102.2 tok/s 62.6 tok/s 102.2 tok/s

响应

我是 Claude，由 Anthropic 开发的 AI 助手。

ææ¯ Claudeï¼ç± Anthropic å¼åç AI å©æãå½åçæ¬ä¸º claude-sonnet-4-6ï¼è´å©ã

ææ¯ Claudeï¼ç± Anthropic æé ç AI å©æãå½åè¿è¡çæ¨¡åçæ¬ä¸º claude-sonnet-4-6ã

ææ¯ Claudeï¼ç± Anthropic å¼åç AI å©æãæè½å¸®å©åçé®é¢ãåè¡çæ¨¡åçæ¬ä¸º claude-sonnet-4-6ã

步骤 5: knowledge_recall text

时间 23:58:15 ~ 23:58:35 (18.4s) 23:56:29 ~ 23:56:44 (14.9s) 23:59:17 ~ 23:59:33 (14.4s) 23:59:53 ~ 07/03 00:00:13 (17.2s)

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) 2,180 ms 1,397 ms 1,359 ms 2,324 ms

首文本(SSE delta) 4,397 ms 1,398 ms 1,360 ms 2,325 ms

总耗时 4,555 ms 3,269 ms 3,273 ms 3,323 ms

输入 3 3 3 1,191

输出 198 204 204 13

缓存创建

6,456

5m: 0 / 1h: 6,456

6,789

5m: 0 / 1h: 6,789

5m: 0 / 1h: 0

1,759

5m: 0 / 1h: 1,759

缓存读取 0 0 6,789 12,100

输出速度 - 109.0 tok/s 106.6 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

{"title": "Knowledge Cutoff Response"}

步骤 6: digit_count text

时间 07/03 00:01:16 ~ 00:01:21 (1.1s) 23:57:53 ~ 23:58:08 (11.7s) 07/03 00:02:12 ~ 00:02:27 (11.8s) -

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=705121164705956112164386054291602131490469446122132653512044195845143073

HTTP 头(TTFB) - 1,066 ms 2,597 ms -

首文本(SSE delta) - 1,072 ms 2,972 ms -

总耗时 - 10,681 ms 10,813 ms -

输入 - 164 164 -

输出 - 1,016 808 -

缓存创建

5m: 0 / 1h: 0

缓存读取 - 0 0 -

输出速度 - 105.7 tok/s 103.0 tok/s -

响应

失败

评测失败

```json
{"n":72,"c":[7,3,5]}
```

**Verification:**

String: `705121164705956112164386054291602131490469446122132653512044195845143073`

- **Length**: 72 digits
- **0s** (×7): positions 2,11,25,32,39,

```json
{"n":72,"c":[7,3,5]}
```

**Verification:**

String: `705121164705956112164386054291602131490469446122132653512044195845143073`

- **Length**: 72 digits
- **0s**: positions 2,11,25,32,39,58,70

未执行

总计

总输入(含缓存)

94,779

仅前 6 步

96,908

96,740

73,814

仅前 5 步

总输出

257

仅前 6 步

1,315

1,091

122

仅前 5 步

总缓存创建

28,634

5m: 0 / 1h: 28,634

仅前 6 步

29,375

5m: 0 / 1h: 29,375

285

5m: 0 / 1h: 285

16,555

5m: 0 / 1h: 16,555

仅前 5 步

总缓存读取

66,130

仅前 6 步

67,354

96,276

56,056

仅前 5 步

输出速度

仅前 6 步

106.1 tok/s

101.0 tok/s

102.2 tok/s

仅前 5 步

总执行时间

29m 30.9s

仅前 6 步

32m 12.3s

36m 16.6s

37m 22.4s

仅前 5 步

总墙钟时间

14m 6.9s

含步骤间等待

仅前 6 步

12m 5.8s

含步骤间等待

13m 13.6s

含步骤间等待

12m 26.3s

含步骤间等待

仅前 5 步

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1311

Anthropic: Claude Sonnet 4.6 (部分)

$0.1506

Anthropic: Claude Sonnet 4.6

$0.0469

Anthropic: Claude Sonnet 4.6

$0.0843

Anthropic: Claude Sonnet 4.6 (部分)

通道指纹

来源步骤 knowledge_recall digit_count digit_count knowledge_recall

平台 direct (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

ID 格式 msg_012UTwhMPRqTVkuo8BwGR7hE msg_01CoPfy2MLzPe4Pdz8efgS8n msg_01QQnYiX3uyu4i977iKURt4S msg_01XNbDY713EyUfTJLMBeb91B

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-L2	cf-ray: a14ed9f44e993511-LAX server: Photon-Edge	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	1783007900684-z99o8vgs6963q req_011CcdWMSRyxQt2KuighNBaw a14ed9f44e993511-LAX
O-Max	server: nginx set-cookie: server_name_session=48b449e29b720b742b1fbf23591a9a4f; Max-Age=86400; httponly; path=/ x-new-api-version: v1.0.0-rc.15 x-oneapi-request-id: 202607021557547708141718268d9d6bW417Q3f	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_011CcdWebBK9dvC9jxZJUmhE
O-Max-NL	via: 1.1 Caddy cf-ray: a14edc5f88d29256-FRA server: cloudflare x-new-api-version: v0.12.10 x-oneapi-request-id: 2026070216000099082608268d9d6WkZ432X4	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: true 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:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a14edc5f88d29256-FRA

协议指纹评分（机器自动）

总分 / 100 98 97 -

基线状态已对照基线已对照基线 -

模型匹配 — 10.0 /10 ×14 10.0 /10 ×14 -

缓存连续性 — 10.0 /10 ×14 10.0 /10 ×14 -

sliding 5m — 10.0 /10 ×13 10.0 /10 ×13 -

缓存 TTL 一致 — 10.0 /10 ×15 10.0 /10 ×15 -

缓存命中比 — 10.0 /10 ×20 10.0 /10 ×20 -

身份(结构) — 10.0 /10 ×7 10.0 /10 ×7 -

知识截止 — 10.0 /10 ×5 10.0 /10 ×5 -

身份(自由) — 10.0 /10 ×7 10.0 /10 ×7 -

原生 msg-ID — 10.0 /10 ×8 10.0 /10 ×8 -

Req-ID 透传 — 0.0 /10 ×2 10.0 /10 ×2 -

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

service_tier — 10.0 /10 ×6 10.0 /10 ×6 -

inference_geo — 10.0 /10 ×5 10.0 /10 ×5 -

SDK 一致 — 10.0 /10 ×2 10.0 /10 ×2 -

系统提示纯净 — 10.0 /10 ×8 10.0 /10 ×8 -

延迟基线 — 10.0 /10 ×15 10.0 /10 ×15 -

流式投递 — 未计入未计入 -

综合结论

与基线相似度执行失败，不评分 98 v3.27.0 基准（参考）未评分（旧数据）

雷达图例

共 19 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 档位思考量 tier_thinking_volume_match
19 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 3970 ms 4450 ms 4530 ms 3410 ms

输出速度 - 106.1 tok/s 101.0 tok/s 102.2 tok/s

完成步骤 5/6 成功 · 1 失败 6/6 成功 6/6 成功 5/6 成功 · 1 跳过

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1311 (部分)

$0.1506

$0.0469

$0.0843 (部分)

缓存读取占比 ⓘ 70% (部分) 70% 100% 76% (部分)

揭晓

通道

服务商 SSSAiCode ↗

通道O-Max-L2

服务商 LinkAPI ↗

通道O-Max

官方基准

服务商 sucui ↗

通道O-Max-NL

返回列表