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。

3 通道 · 6 步骤

2026/06/20 运行日期

claude-opus-4-8 官方基准

2 通道分数 < 70 另含 2 任务失败

已揭晓揭晓进度

⚠ 2 通道不可测 · 已从主对比表移除点击展开原因

AnyAIGC / M-
```
评测失败
```
TopRouterCN / O-Max
```
评测失败
```

步骤 / 指标

官方基准

Novaris

O-Max

luminai

M-Kiro/Antigravity

基本信息

服务名 Claude Code Claude Code Claude Code

模型 claude-opus-4-8 claude-opus-4-8 claude-opus-4-8

客户端 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 20:01:22 ~ 20:01:27 (3.4s) 20:01:13 ~ 20:01:19 (3.9s) 20:01:19 ~ 20:01:30 (5.5s)

提示词

查看本步 Prompt

ping. Reply with the single word: pong

HTTP 头(TTFB) 1,982 ms 2,900 ms 4,189 ms

首文本(SSE delta) 1,983 ms 2,901 ms 4,191 ms

总耗时 2,141 ms 3,101 ms 4,343 ms

输入 1,454 1,492 1,027

输出 4 4 1

缓存创建

30,222

5m: 0 / 1h: 30,222

27,567

5m: 27,567 / 1h: 0

2,925

5m: 2,925 / 1h: 0

缓存读取 0 0 16,576

响应

pong

pong

pong

步骤 2: identity text

时间 20:04:12 ~ 20:04:17 (3.7s) 20:02:43 ~ 20:02:48 (3.7s) 20:03:17 ~ 20:03:37 (17.5s)

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。请只输出以下 3 行内容（label 小写，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) 2,332 ms 2,018 ms 16,292 ms

首文本(SSE delta) 2,333 ms 2,019 ms 16,293 ms

总耗时 2,704 ms 2,761 ms 16,381 ms

输入 2 2 1,035

输出 28 30 23

缓存创建

1,630

5m: 0 / 1h: 1,630

1,668

5m: 1,668 / 1h: 0

2,949

5m: 2,949 / 1h: 0

缓存读取 30,222 27,567 16,715

输出速度 - 40.4 tok/s -

响应

vendor: anthropic
brand: claude
model: claude-opus-4-8

vendor: anthropic
brand: claude
model: claude-opus-4-20250514

vendor: anthropic
brand: claude
model: claude-opus-4-8

步骤 3: cutoff text

时间 20:07:31 ~ 20:07:36 (4.6s) 20:04:14 ~ 20:04:19 (4.2s) 20:07:37 ~ 20:07:47 (6.6s)

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) 3,131 ms 2,481 ms 3,943 ms

首文本(SSE delta) 3,133 ms 2,482 ms 3,944 ms

总耗时 3,470 ms 3,335 ms 4,123 ms

输入 2 2 1,039

输出 6 6 4

缓存创建

5m: 0 / 1h: 64

5m: 66 / 1h: 0

2,958

5m: 2,958 / 1h: 0

缓存读取 31,852 29,235 16,764

响应

2026-01

2025-03

2026-01

步骤 4: identity_free text

时间 20:08:54 ~ 20:08:59 (4.9s) 20:07:39 ~ 20:07:44 (4.3s) 20:11:37 ~ 20:11:47 (6.5s)

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) 2,845 ms 2,293 ms 4,436 ms

首文本(SSE delta) 2,846 ms 2,342 ms 5,000 ms

总耗时 3,787 ms 3,314 ms 5,071 ms

输入 2 2 1,041

输出 58 66 46

缓存创建

5m: 0 / 1h: 37

5m: 39 / 1h: 0

2,964

5m: 2,964 / 1h: 0

缓存读取 31,916 29,301 16,801

输出速度 61.6 tok/s 67.9 tok/s -

响应

我是 Claude，属于 Claude 产品系列，由 Anthropic 公司开发的大型语言模型。具体型号为 Claude Opus 4.8。

我是 Claude Code，由 Anthropic 开发的官方命令行工具，基于 Claude 大语言模型，帮助开发者在终端中完成编程、调试和各类任务。

我是 Claude Opus 4.8，由 Anthropic 开发的 AI 助手，可以帮你完成编程、写作、分析等各类任务。

步骤 5: knowledge_recall text

时间 20:11:03 ~ 20:11:13 (9.4s) 20:09:37 ~ 20:09:47 (7.4s) 20:14:58 ~ 20:15:13 (11.9s)

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) 3,059 ms 1,714 ms 6,635 ms

首文本(SSE delta) 5,924 ms 4,593 ms 6,636 ms

总耗时 8,429 ms 6,728 ms 10,838 ms

输入 2 2 169

输出 654 652 263

缓存创建

4,765

5m: 0 / 1h: 4,765

3,070

5m: 3,070 / 1h: 0

479

5m: 479 / 1h: 0

缓存读取 0 0 2,715

输出速度 burst burst 62.6 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

步骤 6: digit_count text

时间 20:14:31 ~ 20:14:36 (2.4s) 20:11:19 ~ 20:11:29 (8.8s) 20:16:27 ~ 20:16:43 (11.8s)

提示词

查看本步 Prompt

Count digits exactly in s. Return {"n":len(s),"c":[count0,count7,count9]}. s=994335902469050564275932870073174130465870720311564992164048290869005116

HTTP 头(TTFB) 1,110 ms 1,837 ms 6,797 ms

首文本(SSE delta) 1,112 ms 7,806 ms 6,798 ms

总耗时 1,484 ms 8,089 ms 10,497 ms

输入 211 193 188

输出 19 734 83

缓存创建

5m: 0 / 1h: 0

缓存读取 0 0 -

输出速度 - burst 22.4 tok/s

响应

{"n":70,"c":[10,5,9]}

{"n":72,"c":[12,6,9]}

{"n":72,"c":[12,6,9]}

总计

总输入(含缓存) 132,381 120,206 86,345

总输出 769 1,492 420

总缓存创建

36,718

5m: 0 / 1h: 36,718

32,410

5m: 32,410 / 1h: 0

12,275

5m: 12,275 / 1h: 0

总缓存读取 93,990 86,103 69,571

输出速度 61.6 tok/s 56.0 tok/s 43.8 tok/s

总执行时间 28.5s 32.4s 59.8s

总墙钟时间

13m 14.0s

含步骤间等待

10m 15.8s

含步骤间等待

15m 23.1s

含步骤间等待

按官方价目重估

同等内容若直发 Anthropic 的估算 ⓘ

$0.3041

Anthropic: Claude Opus 4.8

$0.2154

Anthropic: Claude Opus 4.8

$0.1157

Anthropic: Claude Opus 4.8

通道指纹

来源步骤 digit_count digit_count digit_count

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

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

CDN - - cloudflare

ID 格式 msg_01V8SYx8StuLvf9cXGqtMmb3 msg_016QBdVdEURvrhoTa1gVjUAw msg_01q6dbaSkkGQLP4vG1V56iusFi

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_011CcEVaZbW6dzo2jvSSUqj3

O-Max

server: nginx
x-oneapi-request-id: 202606201211204685074918268d9d6JV3SuA7u

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

—

M-Kiro/Antigravity

cf-ray: a0eab4775e6b20ec-AMS
server: cloudflare
x-oneapi-request-id: 202606201216294293599968268d9d6UkvRnkut

id_format: msg_prefix
extra_fields: ["container"]
has_service_tier: true
has_inference_geo: false
model_has_date_suffix: false
usage_has_cache_fields: true

hdr:x-oneapi-request-idhdr:server:cloudflarehdr:cf-raybody:id:msg_prefixbody:usage.service_tierbody:usage.cache_fieldsbody:extra_fieldsupstream:anthropic-direct-from-body

a0eab4775e6b20ec-AMS

协议指纹评分（机器自动）

总分 / 100 100 69 74

基线状态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

缓存 TTL 一致 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

缓存连续性 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

原生 msg-ID 10.0 /10 ×8 10.0 /10 ×8 10.0 /10 ×8

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

身份(结构) 10.0 /10 ×7 6.7 /10 ×7 10.0 /10 ×7

知识截止 10.0 /10 ×7 0.0 /10 ×7 10.0 /10 ×7

身份(自由) 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

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

延迟基线 10.0 /10 ×5 10.0 /10 ×5 9.1 /10 ×5

流式投递 10.0 /10 ×5 10.0 /10 ×5 0.0 /10 ×5

Req-ID 透传 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

SDK 一致 10.0 /10 ×2 10.0 /10 ×2 10.0 /10 ×2

综合结论

与基线相似度基准（参考） 69 v3.24.0 74 v3.24.0

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 流式投递 buffer_dump_match
6 自由身份 identity_free_clean
7 停止原因 stop_reason_present
8 系统提示纯净 system_prompt_clean
9 Service Tier service_tier_present
10 缓存命中比 cache_hit_ratio_match
11 缓存 TTL 一致 cache_ttl_consistency
12 推理地区 inference_geo_present
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 信封自报一致 envelope_self_report_match
20 中文遵循 instruction_following_lang
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 3669 ms 4555 ms 8542 ms

输出速度 61.6 tok/s 56.0 tok/s 43.8 tok/s

完成步骤 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 ⓘ

$0.3041

$0.2154

$0.1157

缓存读取占比 ⓘ 71% 72% 81%

揭晓

通道

官方基准

服务商 Novaris ↗

通道O-Max

服务商 luminai ↗

通道M-Kiro/Antigravity

返回列表