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

2026/06/28 运行日期

claude-haiku-4-5-20251001 官方基准

2 通道分数 < 70 与基线相似度

已揭晓揭晓进度

步骤 / 指标

FoxCode

R-super

FoxCode

R-ultra

PackyCode

cc-Zeus

官方基准

TopRouterCN

O-Max

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 22:08:28 ~ 22:08:33 (3.5s) 22:08:30 ~ 22:08:35 (3.7s) 22:08:34 ~ 22:08:39 (4.3s) 22:08:36 ~ 22:08:41 (3.4s) 22:08:32 ~ 22:08:37 (4.6s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 2,180 ms 2,183 ms 1,734 ms 1,652 ms 3,591 ms

首文本(SSE delta) 2,182 ms 2,185 ms 1,736 ms 1,653 ms 3,593 ms

总耗时 2,732 ms 2,798 ms 2,171 ms 2,318 ms 3,729 ms

输入 33,020 33,019 10 10 1,645

输出 51 48 54 53 60

缓存创建

5m: 0 / 1h: 0

33,122

5m: 0 / 1h: 33,122

33,122

5m: 0 / 1h: 33,122

31,487

5m: 0 / 1h: 31,487

缓存读取 0 0 0 0 0

输出速度 92.7 tok/s 78.3 tok/s 124.1 tok/s 79.7 tok/s -

响应

pong

pong

pong

pong

pong

步骤 2: identity text

时间 22:11:18 ~ 22:11:23 (4.5s) 22:12:00 ~ 22:12:10 (5.5s) 22:10:13 ~ 22:10:23 (8.7s) 22:11:20 ~ 22:11:30 (5.8s) 22:12:22 ~ 22:12:32 (5.2s)

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) 1,696 ms 1,992 ms 1,968 ms 2,148 ms 4,340 ms

首文本(SSE delta) 1,697 ms 1,993 ms 1,969 ms 2,150 ms 4,342 ms

总耗时 3,806 ms 4,715 ms 7,013 ms 4,806 ms 4,432 ms

输入 33,576 33,568 10 10 10

输出 242 330 259 250 211

缓存创建

5m: 0 / 1h: 0

238

5m: 0 / 1h: 238

237

5m: 0 / 1h: 237

1,879

5m: 0 / 1h: 1,879

缓存读取 0 0 33,122 33,122 31,487

输出速度 114.7 tok/s 121.2 tok/s 51.3 tok/s 94.1 tok/s burst

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

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

步骤 3: cutoff text

时间 22:14:48 ~ 22:14:54 (3.9s) 22:15:49 ~ 22:15:55 (3.3s) 22:11:48 ~ 22:11:54 (3.3s) 22:12:31 ~ 22:12:37 (3.3s) 22:15:12 ~ 22:15:17 (4.2s)

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) 2,006 ms 1,869 ms 1,370 ms 1,572 ms 3,458 ms

首文本(SSE delta) 2,008 ms 1,870 ms 1,371 ms 1,574 ms 3,459 ms

总耗时 2,829 ms 2,577 ms 2,119 ms 2,369 ms 3,548 ms

输入 34,205 34,226 10 10 10

输出 82 54 75 68 67

缓存创建

5m: 0 / 1h: 0

299

5m: 0 / 1h: 299

273

5m: 0 / 1h: 273

253

5m: 0 / 1h: 253

缓存读取 0 0 33,360 33,359 33,366

输出速度 99.9 tok/s 76.4 tok/s 100.3 tok/s 85.5 tok/s -

响应

2025-02

2025-02

2025-02

2025-02

2025-02

步骤 4: identity_free text

时间 22:16:49 ~ 22:16:54 (4.6s) 22:19:05 ~ 22:19:15 (8.4s) 22:14:11 ~ 22:14:21 (5.6s) 22:16:01 ~ 22:16:16 (14.1s) 22:17:42 ~ 22:17:57 (11.2s)

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) 2,308 ms 2,452 ms 1,420 ms 2,059 ms 8,803 ms

首文本(SSE delta) 2,309 ms 2,454 ms 1,421 ms 2,061 ms 8,804 ms

总耗时 3,785 ms 7,602 ms 3,599 ms 12,983 ms 10,421 ms

输入 34,634 34,620 10 10 10

输出 257 926 354 1,134 741

缓存创建

5m: 0 / 1h: 0

117

5m: 0 / 1h: 117

106

5m: 0 / 1h: 106

105

5m: 0 / 1h: 105

缓存读取 0 0 33,659 33,632 33,619

输出速度 174.1 tok/s 179.9 tok/s 162.5 tok/s 103.8 tok/s burst

响应

我是Claude，由Anthropic公司开发的AI助手。我可以帮助编程、分析、写作等多种任务。

我是Claude，由Anthropic公司打造的AI助手。Haiku 4.5模型，专注快速问题解答和代码协助。

我是Claude，由Anthropic公司开发的AI助手。我可以帮助用户进行代码编写、调试、文件编辑等软件工程任务。

我是Claude Haiku 4.5，由Anthropic公司开发的AI助手。可进行对话、编码、分析和推理等任务。

我是Claude Haiku 4.5，由Anthropic开发。这是Claude系列的轻量高效版本，专为编程辅助、研究和问答设计。

步骤 5: knowledge_recall text

时间 22:18:43 ~ 22:18:53 (9.7s) 22:22:31 ~ 22:22:46 (14.7s) 22:16:55 ~ 22:17:35 (37.5s) 22:20:09 ~ 22:20:24 (12.3s) 22:19:44 ~ 22:19:59 (14.6s)

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) 1,433 ms 1,490 ms 1,508 ms 1,478 ms 7,308 ms

首文本(SSE delta) 1,434 ms 1,491 ms 1,509 ms 1,480 ms 7,310 ms

总耗时 8,788 ms 13,859 ms 36,580 ms 11,363 ms 13,808 ms

输入 10,218 10,897 10 10 10

输出 1,439 2,197 4,964 1,309 1,845

缓存创建

5m: 0 / 1h: 0

8,748

5m: 0 / 1h: 8,748

9,490

5m: 0 / 1h: 9,490

9,083

5m: 0 / 1h: 9,083

缓存读取 0 0 0 0 0

输出速度 195.7 tok/s 177.6 tok/s 141.5 tok/s 132.4 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

{"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

```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 22:20:35 ~ 22:20:45 (7.9s) 22:25:11 ~ 22:25:31 (16.9s) 22:19:31 ~ 22:20:01 (26.0s) 22:24:18 ~ 22:24:53 (34.4s) 22:22:29 ~ 22:22:54 (24.6s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=519083228538616023723241144635843609030365194969979247478984438390015357

HTTP 头(TTFB) 1,266 ms 1,160 ms 766 ms 989 ms 6,223 ms

首文本(SSE delta) 1,267 ms 1,161 ms 767 ms 991 ms 6,225 ms

总耗时 7,001 ms 16,111 ms 24,771 ms 33,486 ms 23,944 ms

输入 209 209 193 193 193

输出 1,401 3,371 3,855 4,344 2,717

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 0 0 0

输出速度 244.3 tok/s 225.5 tok/s 160.6 tok/s 133.7 tok/s 153.3 tok/s

响应

```json
{"n":72,"c":[7,5,9]}
```

```json
{"n":72,"c":[7,5,9]}
```

```json
{"n":72,"c":[7,5,9]}
```

```json
{"n":72,"c":[7,5,9]}
```

```json
{"n":72,"c":[7,5,9]}
```

总计

总输入(含缓存) 145,862 146,539 142,908 143,584 143,157

总输出 3,472 6,926 9,561 7,158 5,641

总缓存创建

5m: 0 / 1h: 0

42,524

5m: 0 / 1h: 42,524

43,228

5m: 0 / 1h: 43,228

42,807

5m: 0 / 1h: 42,807

总缓存读取 0 0 100,141 100,113 98,472

输出速度 192.4 tok/s 189.7 tok/s 141.7 tok/s 124.7 tok/s 153.3 tok/s

总执行时间 34.0s 52.6s 1m 25.4s 1m 13.4s 1m 4.5s

总墙钟时间

12m 17.0s

含步骤间等待

17m 1.4s

含步骤间等待

11m 27.0s

含步骤间等待

16m 17.1s

含步骤间等待

14m 22.0s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.1632

Anthropic: Claude Haiku 4.5

$0.1812

Anthropic: Claude Haiku 4.5

$0.1112

Anthropic: Claude Haiku 4.5

$0.1001

Anthropic: Claude Haiku 4.5

$0.0934

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%) direct (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 - -

ID 格式 msg_01epKNmESdHSIVsFkLf7xfjq msg_01cniYd08eFuoNlj1mNp2EbP msg_01PgzDGyCQCcnVbTX8Hcndv6 msg_01Kj2NgWC9n3N8qeBeGR7f1t msg_01GHkGRjXxLCkFzGEui3Y96N

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
R-super	cf-ray: a12d5547b879e041-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606281420365959815118268d9d6QyJSXA3S	id_format: msg_prefix 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-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	a12d5547b879e041-NRT
R-ultra	cf-ray: a12d5c0609afe041-NRT server: cloudflare x-new-api-version: v0.0.0 x-oneapi-request-id: 202606281425128105249168268d9d6ablA1JrP	id_format: msg_prefix 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-raybody:id:msg_prefixbody:usage.service_tierbody:usage.inference_geobody:model:date_suffixbody:usage.cache_fieldsupstream:anthropic-direct-from-body	a12d5c0609afe041-NRT
cc-Zeus	via: 1.1 Caddy cf-ray: a12d53ba08cb2f7c-LAX server: cloudflare x-oneapi-request-id: 20260628221932825298370g324mrI0	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-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	a12d53ba08cb2f7c-LAX
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_011CcVowwZ69be9hmNYLEsB9
O-Max	server: nginx/1.27.5	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	tr-ew7y4grd-k7ftlc0abs-139429

协议指纹评分（机器自动）

总分 / 100 66 67 98 100 97

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 0.0 /10 ×20 0.0 /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 5.0 /10 ×14 5.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14 10.0 /10 ×14

sliding 5m 5.0 /10 ×13 5.0 /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 10.0 /10 ×8

身份(结构) 10.0 /10 ×7 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 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 7.7 /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 66 v3.25.0 67 v3.25.0 98 v3.25.0 基准（参考） 97 v3.25.0

雷达图例

共 22 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 档位思考量 tier_thinking_volume_match
22 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 4824 ms 7944 ms 12709 ms 11221 ms 9980 ms

输出速度 192.4 tok/s 189.7 tok/s 141.7 tok/s 124.7 tok/s 153.3 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 ⓘ

$0.1632

$0.1812

$0.1112

$0.1001

$0.0934

缓存读取占比 ⓘ 0% 0% 70% 70% 69%

揭晓

通道

服务商 FoxCode ↗

通道R-super

服务商 FoxCode ↗

通道R-ultra

服务商 PackyCode ↗

通道cc-Zeus

官方基准

服务商 TopRouterCN ↗

通道O-Max

返回列表