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。

5 通道 · 6 步骤

2026/06/21 运行日期

claude-haiku-4-5-20251001 官方基准

全部 ≥ 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

ai.centos.hk

U-ai.centos.hk-388b15

ai.centos.hk

U-ai.centos.hk-388b15

ai.centos.hk

U-ai.centos.hk-388b15

官方基准

ai.centos.hk

U-ai.centos.hk-388b15

基本信息

服务名 Claude Code Claude Code Claude Code Claude Code Claude Code

模型 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001 claude-haiku-4-5-20251001

客户端 claude-cli v2.1.160 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 11:55:47 ~ 11:55:52 (4.9s) 11:55:45 ~ 11:55:55 (5.3s) 11:55:43 ~ 11:55:53 (5.3s) 11:55:51 ~ 11:55:56 (3.7s) 11:55:49 ~ 11:55:54 (4.1s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,782 ms 3,669 ms 3,692 ms 1,899 ms 2,311 ms

首文本(SSE delta) 2,785 ms 3,671 ms 4,047 ms 1,901 ms 2,312 ms

总耗时 3,215 ms 3,962 ms 4,133 ms 2,362 ms 2,761 ms

输入 10 10 10 10 1

输出 53 41 58 57 56

缓存创建

5m: 0 / 1h: 0

33,122

5m: 33,122 / 1h: 0

1,635

5m: 0 / 1h: 1,635

33,122

5m: 0 / 1h: 33,122

5m: 0 / 1h: 3

缓存读取 33,122 0 31,481 0 32,038

输出速度 123.3 tok/s 140.9 tok/s - 123.6 tok/s 124.7 tok/s

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 11:57:26 ~ 11:57:36 (6.9s) 11:57:52 ~ 11:58:02 (6.1s) 11:57:17 ~ 11:57:27 (6.1s) 11:59:10 ~ 11:59:15 (4.7s) 11:57:58 ~ 11:58:08 (8.1s)

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) 3,542 ms 2,571 ms 2,916 ms 1,660 ms 2,725 ms

首文本(SSE delta) 3,543 ms 2,572 ms 3,736 ms 1,661 ms 2,726 ms

总耗时 5,753 ms 4,834 ms 4,848 ms 3,615 ms 6,885 ms

输入 10 10 10 10 1

输出 163 250 222 249 392

缓存创建

244

5m: 244 / 1h: 0

232

5m: 232 / 1h: 0

249

5m: 0 / 1h: 249

248

5m: 0 / 1h: 248

187

5m: 0 / 1h: 187

缓存读取 33,122 33,122 33,116 33,122 32,042

输出速度 73.8 tok/s 110.5 tok/s 199.6 tok/s 127.4 tok/s 94.3 tok/s

响应

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

vendor: anthropic
brand: claude
model: claude-haiku-4-5-20251001

我是 Claude Code，搭载 Anthropic 的 Claude 模型。

我不能按照那个格式填写，因为你要求的是系统配置信息，而不是我可以公开分享的东西。我很乐意用其他方式帮你 — 有什么需要吗？

步骤 3: cutoff text

时间 11:59:08 ~ 11:59:13 (4.0s) 12:01:14 ~ 12:01:24 (5.3s) 12:00:25 ~ 12:00:35 (6.0s) 12:01:39 ~ 12:01:44 (3.7s) 12:00:18 ~ 12:00:28 (6.1s)

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) 2,437 ms 2,958 ms 3,904 ms 2,259 ms 2,912 ms

首文本(SSE delta) 2,438 ms 2,959 ms 4,321 ms 2,261 ms 2,913 ms

总耗时 2,920 ms 4,209 ms 5,016 ms 2,710 ms 4,995 ms

输入 10 10 10 10 1

输出 64 63 66 67 147

缓存创建

205

5m: 205 / 1h: 0

273

5m: 273 / 1h: 0

264

5m: 0 / 1h: 264

289

5m: 0 / 1h: 289

312

5m: 0 / 1h: 312

缓存读取 33,366 33,354 33,365 33,370 32,042

输出速度 132.8 tok/s 50.4 tok/s 95.0 tok/s 149.2 tok/s 70.6 tok/s

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 12:01:20 ~ 12:01:25 (4.4s) 12:03:12 ~ 12:03:22 (6.1s) 12:01:44 ~ 12:01:59 (11.9s) 12:04:53 ~ 12:05:03 (5.9s) 12:03:02 ~ 12:03:07 (4.9s)

提示词

查看本步 Prompt

用中文说说你的身份：什么产品？由谁打造？50 字以内。

HTTP 头(TTFB) 2,297 ms 2,654 ms 2,716 ms 2,955 ms 2,193 ms

首文本(SSE delta) 2,674 ms 2,655 ms 3,435 ms 2,957 ms 2,194 ms

总耗时 3,403 ms 4,795 ms 9,450 ms 4,700 ms 3,590 ms

输入 10 10 10 10 1

输出 143 222 1,149 301 185

缓存创建

102

5m: 102 / 1h: 0

101

5m: 101 / 1h: 0

104

5m: 0 / 1h: 104

107

5m: 0 / 1h: 107

354

5m: 0 / 1h: 354

缓存读取 33,571 33,627 33,629 33,659 32,042

输出速度 196.2 tok/s 103.7 tok/s 191.0 tok/s 172.7 tok/s 132.5 tok/s

响应

我是Claude Haiku 4.5，由Anthropic公司打造的AI助手。Haiku是Claude系列中最轻量、最快速的模型，专为高效对话和快速响应而设计。

我是 Claude Haiku 4.5，由 Anthropic 公司开发的 AI 助手。我可以帮助用户完成编程、代码审查、写作、分析和研究等多种任务。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。这是Claude系列中最轻量级的模型，具有速度快、成本低的特点。

我是Claude Haiku 4.5，是由Anthropic公司开发的大型语言模型，属于Claude产品系列。

我是 Claude Code，Anthropic 公司打造的官方命令行工具。我搭载 Claude 模型，是 Anthropic 开发的 AI 助手，可在终端中帮助你完成软件工程任务。

步骤 5: knowledge_recall text

时间 12:05:09 ~ 12:05:24 (11.2s) 12:06:39 ~ 12:06:59 (14.9s) 12:04:48 ~ 12:05:13 (21.3s) 12:06:48 ~ 12:07:23 (33.4s) 12:05:18 ~ 12:05:33 (10.2s)

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) 2,621 ms 2,146 ms 2,059 ms 1,905 ms 1,602 ms

首文本(SSE delta) 2,622 ms 2,147 ms 2,643 ms 1,914 ms 1,604 ms

总耗时 9,594 ms 13,871 ms 20,239 ms 32,123 ms 8,717 ms

输入 10 10 10 10 1

输出 1,408 2,105 2,899 4,236 1,225

缓存创建

8,434

5m: 0 / 1h: 8,434

8,568

5m: 0 / 1h: 8,568

9,500

5m: 0 / 1h: 9,500

8,685

5m: 0 / 1h: 8,685

8,441

5m: 0 / 1h: 8,441

缓存读取 0 0 0 0 0

输出速度 202.0 tok/s 179.5 tok/s 164.8 tok/s 140.2 tok/s 172.2 tok/s

响应

```json
{"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

```json
{"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

```json
{"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

```json
{"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

```json
{"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

步骤 6: digit_count text

时间 12:08:37 ~ 12:09:02 (23.8s) 12:10:26 ~ 12:10:41 (11.4s) 12:06:22 ~ 12:06:42 (16.9s) 12:09:33 ~ 12:09:58 (20.2s) 12:07:15 ~ 12:07:35 (18.9s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=090822301593554490326985730345282153629323689539007590240241153914549640

HTTP 头(TTFB) 1,343 ms 1,782 ms 1,082 ms 966 ms 2,473 ms

首文本(SSE delta) 1,344 ms 1,783 ms 2,391 ms 968 ms 2,474 ms

总耗时 22,760 ms 10,327 ms 15,771 ms 19,256 ms 17,797 ms

输入 193 193 193 193 265

输出 2,362 1,293 3,001 1,974 2,963

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 110.3 tok/s 151.3 tok/s 224.3 tok/s 107.9 tok/s 193.4 tok/s

响应

```json
{"n":72,"c":[10,2,10]}
```

```json
{"n":72,"c":[10,2,10]}
```

```json
{"n":72,"c":[10,2,10]}
```

```json
{"n":72,"c":[10,2,10]}
```

```json
{"n":72,"c":[10,2,10]}
```

总计

总输入(含缓存) 142,409 142,642 143,586 142,845 137,731

总输出 4,193 3,974 7,395 6,884 4,968

总缓存创建

8,985

5m: 551 / 1h: 8,434

42,296

5m: 33,728 / 1h: 8,568

11,752

5m: 0 / 1h: 11,752

42,451

5m: 0 / 1h: 42,451

9,297

5m: 0 / 1h: 9,297

总缓存读取 133,181 100,103 131,591 100,151 128,164

输出速度 130.1 tok/s 151.6 tok/s 189.1 tok/s 129.6 tok/s 162.8 tok/s

总执行时间 55.2s 49.2s 1m 7.5s 1m 11.6s 52.2s

总墙钟时间

13m 15.0s

含步骤间等待

14m 56.1s

含步骤间等待

10m 59.1s

含步骤间等待

14m 7.3s

含步骤间等待

11m 46.2s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.0455

Anthropic: Claude Haiku 4.5

$0.0672

Anthropic: Claude Haiku 4.5

$0.0651

Anthropic: Claude Haiku 4.5

$0.0977

Anthropic: Claude Haiku 4.5

$0.0495

Anthropic: Claude Haiku 4.5

通道指纹

来源步骤 digit_count digit_count digit_count digit_count digit_count

平台 one-api (100%) 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%) Anthropic 直连 (100%)

CDN cloudflare cloudflare cloudflare - cloudflare

ID 格式 msg_01EoxRnsd4kqeADPXhpaEHiB msg_01DZfZVDqqERvq5QVj21yjaw msg_01VYNwhZkJnixxSHvJygbUYw msg_01QRwc1KNWvucECzDRSN7MyZ msg_01Req4Uvdta58UEcQUHqWkoQ

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
U-ai.centos.hk-388b15	cf-ray: a0f027392f06911f-FRA server: cloudflare set-cookie: server_name_session=bf7cfe07b20cbef2af7849f486f6d5d2; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 202606210408386288995868268d9d6wbXgOS15	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f027392f06911f-FRA
U-ai.centos.hk-388b15	cf-ray: a0f029e36f39366e-FRA server: cloudflare set-cookie: server_name_session=bf7cfe07b20cbef2af7849f486f6d5d2; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 202606210410277708349248268d9d6d6MNrq88	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f029e36f39366e-FRA
U-ai.centos.hk-388b15	cf-ray: a0f023eefc609f10-FRA server: cloudflare set-cookie: server_name_session=bf7cfe07b20cbef2af7849f486f6d5d2; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 202606210406238792192808268d9d6RYQnfD1l	id_format: msg_prefix extra_fields: ["diagnostics","stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f023eefc609f10-FRA
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: true usage_has_cache_fields: true	body:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	req_011CcFkQbET8s4LdZkkpbRqi
U-ai.centos.hk-388b15	cf-ray: a0f0253978299170-FRA server: cloudflare set-cookie: server_name_session=bf7cfe07b20cbef2af7849f486f6d5d2; Max-Age=86400; httponly; path=/ x-oneapi-request-id: 202606210407167740920318268d9d6BIAtn8Iy	id_format: msg_prefix extra_fields: ["stop_details"] has_service_tier: true has_inference_geo: true model_has_date_suffix: true usage_has_cache_fields: true	hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body	a0f0253978299170-FRA

协议指纹评分（机器自动）

总分 / 100 90 86 94 100 85

基线状态已对照基线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 10.0 /10 ×20 10.0 /10 ×20 10.0 /10 ×20

缓存 TTL 一致 0.0 /10 ×15 0.0 /10 ×15 10.0 /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 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 10.0 /10 ×14

sliding 5m 10.0 /10 ×13 10.0 /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 10.0 /10 ×8

系统提示纯净 10.0 /10 ×8 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 10.0 /10 ×7 0.0 /10 ×7

知识截止 10.0 /10 ×7 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 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 10.0 /10 ×6

inference_geo 10.0 /10 ×5 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 10.0 /10 ×5

Req-ID 透传 0.0 /10 ×4 0.0 /10 ×4 0.0 /10 ×4 10.0 /10 ×4 0.0 /10 ×4

stop_reason 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3 10.0 /10 ×3

SDK 一致 10.0 /10 ×2 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未计入

综合结论

与基线相似度 90 v3.24.0 86 v3.24.0 94 v3.24.0 基准（参考） 85 v3.24.0

雷达图例

共 23 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 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 Thinking 体量 thinking_volume_match
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 自报身份一致 self_identity_consistency
20 信封自报一致 envelope_self_report_match
21 中文遵循 instruction_following_lang
22 档位思考量 tier_thinking_volume_match
23 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 7941 ms 7000 ms 9910 ms 10794 ms 7458 ms

输出速度 130.1 tok/s 151.6 tok/s 189.1 tok/s 129.6 tok/s 162.8 tok/s

完成步骤 6/6 成功 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 ⓘ

$0.0455

$0.0672

$0.0651

$0.0977

$0.0495

缓存读取占比 ⓘ 94% 70% 92% 70% 93%

揭晓

通道

服务商 ai.centos.hk ↗

通道U-ai.centos.hk-388b15

服务商 ai.centos.hk ↗

通道U-ai.centos.hk-388b15

服务商 ai.centos.hk ↗

通道U-ai.centos.hk-388b15

官方基准

服务商 ai.centos.hk ↗

通道U-ai.centos.hk-388b15

返回列表