← 返回 API Dashboard

本机回环 API · 本页为 registry 静态快照

Free Model Chain — 模型分数全览

展示全部免费模型的分层评分与 routing_score 加权计算过程。数据来自 model-registry-latest.json，非实时调用 8022 API。API 仍仅本机 127.0.0.1:8022 可访问。

更新时间

2026-08-10 18:10:50 CST

scoring_version

v2-routing

quality 覆盖

13 / 17 模型已 quality 评测（0.7647）

数据来源

reports/run-20260606-011232/benchmark.json + current OpenRouter free catalog filter + quality benchmarks merged + newer partial gate runs merged

#1

poolside/laguna-s-2.1:free

0.9423 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.9550	×0.45	0.4297
gate_score	1.0000	×0.20	0.2000
stability_score	0.9000	×0.20	0.1800
format_score	1.0000	×0.10	0.1000
latency_score (11330 ms)	0.6500	×0.05	0.0325
合计 → routing_score			0.9422

② Gate 层（smoke test）

gate_score = 1.0000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	1.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.9550
effective_quality_score	0.9550
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 1.0000
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 11330
• BuildSignal recommended: 推荐

#2

google/gemma-4-26b-a4b-it:free

0.9310 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.9600	×0.45	0.4320
gate_score	0.9375	×0.20	0.1875
stability_score	0.9000	×0.20	0.1800
format_score	0.9900	×0.10	0.0990
latency_score (11762 ms)	0.6500	×0.05	0.0325
合计 → routing_score			0.9310

② Gate 层（smoke test）

gate_score = 0.9375（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.7500

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.9600
effective_quality_score	0.9600
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.9500
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 11762
• BuildSignal recommended: 推荐

#3

nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free

0.8879 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.8847	×0.45	0.3981
gate_score	1.0000	×0.20	0.2000
stability_score	0.9000	×0.20	0.1800
format_score	0.8980	×0.10	0.0898
latency_score (42857 ms)	0.4000	×0.05	0.0200
合计 → routing_score			0.8879

② Gate 层（smoke test）

gate_score = 1.0000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	1.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.8847
effective_quality_score	0.8847
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.4900
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 42857
• BuildSignal recommended: 推荐

#4

openrouter/free

0.8704 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.8480	×0.45	0.3816
gate_score	0.9375	×0.20	0.1875
stability_score	0.9000	×0.20	0.1800
format_score	0.8880	×0.10	0.0888
latency_score (15945 ms)	0.6500	×0.05	0.0325
合计 → routing_score			0.8704

② Gate 层（smoke test）

gate_score = 0.9375（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.7500

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.8480
effective_quality_score	0.8480
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.4400
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 15945
• BuildSignal recommended: 推荐

#5

nvidia/nemotron-3-ultra-550b-a55b:free

0.8387 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.7997	×0.45	0.3599
gate_score	1.0000	×0.20	0.2000
stability_score	0.9000	×0.20	0.1800
format_score	0.7880	×0.10	0.0788
latency_score (22029 ms)	0.4000	×0.05	0.0200
合计 → routing_score			0.8387

② Gate 层（smoke test）

gate_score = 1.0000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	1.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.7997
effective_quality_score	0.7997
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.4400
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 22029
• BuildSignal recommended: 推荐

#6

nvidia/nemotron-3-super-120b-a12b:free

0.8166 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.8063	×0.45	0.3628
gate_score	0.8750	×0.20	0.1750
stability_score	0.9000	×0.20	0.1800
format_score	0.7880	×0.10	0.0788
latency_score (47176 ms)	0.4000	×0.05	0.0200
合计 → routing_score			0.8166

② Gate 层（smoke test）

gate_score = 0.8750（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.5000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.8063
effective_quality_score	0.8063
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.4400
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 47176
• BuildSignal recommended: —

#7

nvidia/nemotron-3-nano-30b-a3b:free

0.8084 routing_score 推荐

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.6880	×0.45	0.3096
gate_score	0.9375	×0.20	0.1875
stability_score	0.9000	×0.20	0.1800
format_score	0.8880	×0.10	0.0888
latency_score (5577 ms)	0.8500	×0.05	0.0425
合计 → routing_score			0.8084

② Gate 层（smoke test）

gate_score = 0.9375（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.7500

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.6880
effective_quality_score	0.6880
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.4400
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 5577
• BuildSignal recommended: 推荐

#8

openai/gpt-oss-20b:free

0.7356 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.6953	×0.45	0.3129
gate_score	0.9375	×0.20	0.1875
stability_score	0.7000	×0.20	0.1400
format_score	0.7520	×0.10	0.0752
latency_score (21236 ms)	0.4000	×0.05	0.0200
合计 → routing_score			0.7356

② Gate 层（smoke test）

gate_score = 0.9375（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.7500

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.6953
effective_quality_score	0.6953
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.2600
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 21236
• BuildSignal recommended: —

#9

google/gemma-4-31b-it:free

0.7265 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.9867	×0.45	0.4440
gate_score	0.0000	×0.20	0.0000
stability_score	0.7000	×0.20	0.1400
format_score	1.0000	×0.10	0.1000
latency_score (8363 ms)	0.8500	×0.05	0.0425
合计 → routing_score			0.7265

② Gate 层（smoke test）

gate_score = 0.0000（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	0.0000
简单推理 reasoning	0.0000
BuildSignal 片段	0.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.9867
effective_quality_score	0.9867
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 1.0000
reasoning_score (legacy)	0.0000 = gate 推理

④ 其它

• latency_ms: 8363
• BuildSignal recommended: —

#10

cohere/north-mini-code:free

0.6718 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.5703	×0.45	0.2566
gate_score	1.0000	×0.20	0.2000
stability_score	0.7000	×0.20	0.1400
format_score	0.5520	×0.10	0.0552
latency_score (28583 ms)	0.4000	×0.05	0.0200
合计 → routing_score			0.6718

② Gate 层（smoke test）

gate_score = 1.0000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	1.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.5703
effective_quality_score	0.5703
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.2600
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 28583
• BuildSignal recommended: —

#11

nvidia/nemotron-nano-12b-v2-vl:free

0.6670 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.6933	×0.45	0.3120
gate_score	0.5000	×0.20	0.1000
stability_score	0.7000	×0.20	0.1400
format_score	0.8250	×0.10	0.0825
latency_score (12061 ms)	0.6500	×0.05	0.0325
合计 → routing_score			0.6670

② Gate 层（smoke test）

gate_score = 0.5000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	0.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.6933
effective_quality_score	0.6933
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 1.0000
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 12061
• BuildSignal recommended: —

#12

nvidia/nemotron-nano-9b-v2:free

0.5676 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.4887	×0.45	0.2199
gate_score	0.5000	×0.20	0.1000
stability_score	0.7000	×0.20	0.1400
format_score	0.7520	×0.10	0.0752
latency_score (17797 ms)	0.6500	×0.05	0.0325
合计 → routing_score			0.5676

② Gate 层（smoke test）

gate_score = 0.5000（四题通过率均值）

中文 zh_summary	1.0000
JSON json_struct	1.0000
简单推理 reasoning	0.0000
BuildSignal 片段	0.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.4887
effective_quality_score	0.4887
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.2600
reasoning_score (legacy)	0.0000 = gate 推理

④ 其它

• latency_ms: 17797
• BuildSignal recommended: —

#13

inclusionai/ling-3.0-tiny:free

0.5200 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.4000	×0.45	0.1800
gate_score	0.5000	×0.20	0.1000
stability_score	0.7000	×0.20	0.1400
format_score	0.5000	×0.10	0.0500
latency_score (4019 ms)	1.0000	×0.05	0.0500
合计 → routing_score			0.5200

② Gate 层（smoke test）

gate_score = 0.5000（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	1.0000
简单推理 reasoning	1.0000
BuildSignal 片段	0.0000

③ Quality 层

未跑 quality benchmark：quality_score 为 gate 兜底；effective = gate − penalty (0.5000 − 0.1000)。quality_score_missing; used gate_score as fallback

quality_score	0.5000
effective_quality_score	0.4000
quality_score_source	gate_fallback
quality_penalty	0.1000
quality_evaluated	否
quality_reasoning: 未评测
reasoning_score (legacy)	1.0000 = gate 推理

④ 其它

• latency_ms: 4019
• BuildSignal recommended: —

#14

nvidia/nemotron-3.5-content-safety:free

0.3398 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.2270	×0.45	0.1022
gate_score	0.0000	×0.20	0.0000
stability_score	0.7000	×0.20	0.1400
format_score	0.5520	×0.10	0.0552
latency_score (6491 ms)	0.8500	×0.05	0.0425
合计 → routing_score			0.3399

② Gate 层（smoke test）

gate_score = 0.0000（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	0.0000
简单推理 reasoning	0.0000
BuildSignal 片段	0.0000

③ Quality 层

quality_score 来自 5 道 quality 题平均分（见仓库 data/README.md）。

quality_score	0.2270
effective_quality_score	0.2270
quality_score_source	quality_benchmark
quality_penalty	0.0000
quality_evaluated	是
quality_reasoning（多约束推理）: 0.2600
reasoning_score (legacy)	0.0000 = gate 推理

④ 其它

• latency_ms: 6491
• BuildSignal recommended: —

#15

poolside/laguna-xs-2.1:free

0.2263 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.0250	×0.45	0.0113
gate_score	0.1250	×0.20	0.0250
stability_score	0.7000	×0.20	0.1400
format_score	0.0000	×0.10	0.0000
latency_score (1041 ms)	1.0000	×0.05	0.0500
合计 → routing_score			0.2263

② Gate 层（smoke test）

gate_score = 0.1250（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	0.0000
简单推理 reasoning	0.5000
BuildSignal 片段	0.0000

③ Quality 层

未跑 quality benchmark：quality_score 为 gate 兜底；effective = gate − penalty (0.1250 − 0.1000)。quality_score_missing; used gate_score as fallback

quality_score	0.1250
effective_quality_score	0.0250
quality_score_source	gate_fallback
quality_penalty	0.1000
quality_evaluated	否
quality_reasoning: 未评测
reasoning_score (legacy)	0.5000 = gate 推理

④ 其它

• latency_ms: 1041
• BuildSignal recommended: —

#16

google/lyria-3-clip-preview

0.1400 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.0000	×0.45	0.0000
gate_score	0.0000	×0.20	0.0000
stability_score	0.7000	×0.20	0.1400
format_score	0.0000	×0.10	0.0000
latency_score (0 ms)	0.0000	×0.05	0.0000
合计 → routing_score			0.1400

② Gate 层（smoke test）

gate_score = 0.0000（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	0.0000
简单推理 reasoning	0.0000
BuildSignal 片段	0.0000

③ Quality 层

未跑 quality benchmark：quality_score 为 gate 兜底；effective = gate − penalty (0.0000 − 0.1000)。quality_score_missing; used gate_score as fallback

quality_score	0.0000
effective_quality_score	0.0000
quality_score_source	gate_fallback
quality_penalty	0.1000
quality_evaluated	否
quality_reasoning: 未评测
reasoning_score (legacy)	0.0000 = gate 推理

④ 其它

• latency_ms: 0
• BuildSignal recommended: —

#17

google/lyria-3-pro-preview

0.1400 routing_score —

① 最终分 routing_score 计算

routing = effective_quality×0.45 + gate×0.20 + stability×0.20 + format×0.10 + latency_score×0.05

项	原始分	×权重	加权贡献
effective_quality_score	0.0000	×0.45	0.0000
gate_score	0.0000	×0.20	0.0000
stability_score	0.7000	×0.20	0.1400
format_score	0.0000	×0.10	0.0000
latency_score (0 ms)	0.0000	×0.05	0.0000
合计 → routing_score			0.1400

② Gate 层（smoke test）

gate_score = 0.0000（四题通过率均值）

中文 zh_summary	0.0000
JSON json_struct	0.0000
简单推理 reasoning	0.0000
BuildSignal 片段	0.0000

③ Quality 层

未跑 quality benchmark：quality_score 为 gate 兜底；effective = gate − penalty (0.0000 − 0.1000)。quality_score_missing; used gate_score as fallback

quality_score	0.0000
effective_quality_score	0.0000
quality_score_source	gate_fallback
quality_penalty	0.1000
quality_evaluated	否
quality_reasoning: 未评测
reasoning_score (legacy)	0.0000 = gate 推理

④ 其它

• latency_ms: 0
• BuildSignal recommended: —

评分细则见仓库 00-projects/22-openrouter-free-benchmark/data/README.md。本页随 build_model_registry.py 自动刷新。