AI 模型评测目录

OPEN BENCHMARKS长期追踪目录

追踪模型在逻辑、数学、规划与人类直觉等私有题库上的长期表现。这一页按报告期次列出该类别的全部成绩，分数、成本与数据日期都标在表里，方便在同一类里挑更稳或更省的模型。三类评测（推理 / Agentic / 视觉）分数互不换算，别跨类比较名次。

收录规模 31 个数据集 当前 44 条模型记录 · 按类别与数据集切换

类别：推理 数据集：2026-10 · 月榜 记录数：44 该类别数据集：31 个


GPT-6 Astra (xhigh) 推理 OpenAI	89.99	86.53	3.84%	+25.1%	312	13091	¥126.46	¥345.00	26-09-05
Opus 5.5 (xhigh ~Estimated) 推理 Anthropic Claude	81.30	77.69	4.44%	+26.3%	295	33096	¥127.88	¥138.00	26-09-23
Fable 5.1 (xhigh ~Estimated) 推理 Anthropic Claude	76.43	70.47	7.80%	-	481	41434	¥400.25	¥345.00	26-09-02
Gemini 4 Argon (high) 推理 Google Gemini	72.62	69.13	4.81%	+27.1%	451	24395	¥47.13	¥69.00	26-10-14
GPT-6.1 Sol (xhigh) 推理 OpenAI	76.16	67.79	10.99%	+23.9%	604	12955	¥25.03	¥69.00	26-09-30
Opus 5.5 (xhigh w/ refuse) 推理 Anthropic Claude	70.24	66.63	5.14%	-	306	31069	¥120.05	¥138.00	26-09-23
Kimi-K3.1-Pro (max) 推理月之暗面 Kimi	72.90	61.65	15.43%	+16%	1316	50846	¥96.81	¥68.00	26-10-24
Muse Wonder 1.0 (xhigh) 推理 Meta	66.21	59.64	9.92%	+33.7%	632	54142	¥209.20	¥138.00	26-10-13
Fable 5.1 (xhigh w/ refuse) 推理 Anthropic Claude	64.65	58.69	9.22%	-	469	37751	¥364.67	¥345.00	26-09-02
Sonnet 5.5 (xhigh) 推理 Anthropic Claude	64.02	57.41	10.32%	+210.7%	364	39996	¥77.27	¥69.00	26-09-29
DeepSeek V4.1 Flash 1017 (max) 推理 DeepSeek 深度求索	62.11	56.03	9.79%	+5.4%	592	76670	¥17.17	¥8.00	26-10-17
Kimi-K3 (max) 推理月之暗面 Kimi	63.86	53.16	16.76%	+32.6%	1487	42024	¥123.55	¥105.00	26-07-16
DeepSeek V4.1 Flash (max) 推理 DeepSeek 深度求索	60.04	52.98	11.76%	+18.9%	634	88921	¥19.92	¥8.00	26-09-10
Grok 4.7 (xhigh) 推理 xAI Grok	63.79	52.55	17.62%	+21.9%	1463	81085	¥75.19	¥33.12	26-09-22
MiniMax-M3.1-Flash (max) 推理 MiniMax 海螺	61.50	51.76	15.84%	+216.9%	719	53276	¥5.60	¥4.00	26-10-19
Gemini 3.8 Flash (high) 推理 Google Gemini	62.06	51.25	17.42%	+9.4%	388	37448	¥27.13	¥25.88	26-09-03
Seed-2.1-pro 0915 (high) 推理字节豆包（火山方舟）	61.41	49.72	19.04%	+44.8%	1443	45900	¥38.56	¥30.00	26-09-16
GLM-5.3 (max) 推理 Z.AI（智谱国际版）	59.12	47.72	19.28%	+16.8%	1553	67573	¥52.98	¥28.00	26-08-19
DeepSeek V4 Pro 0813 (max) 推理 DeepSeek 深度求索	56.02	47.43	15.33%	+26.4%	1707	62275	¥47.08	¥27.00	26-08-13
Qwen3.8-Max (xhigh) 推理阿里云百炼（通义千问）	51.27	45.28	11.68%	-	1646	72215	¥72.79	¥36.00	26-08-03
GLM-5.3-Flash (max) 推理 Z.AI（智谱国际版）	55.50	44.59	19.66%	-	1303	50977	¥4.00	¥2.80	26-08-26
Step 5 Preview 推理阶跃星辰 StepFun	55.30	41.30	25.32%	+413.8%	1326	54501	¥30.52	¥20.00	26-09-20
Qwen3.8-Flash (xhigh) 推理阿里云百炼（通义千问）	50.38	37.57	25.43%	-	1027	67801	¥5.13	¥2.70	26-08-26
Haiku 5.5 (xhigh) 推理 Anthropic Claude	42.82	32.33	24.50%	-	321	59010	¥5.70	¥3.45	26-10-08
Spark X2.6 推理科大讯飞星火	37.58	32.15	14.45%	+253%	1034	74440	¥7.50	¥3.60	26-10-22
Qwen3.8-27B (xhigh) 推理阿里云百炼（通义千问）	41.24	31.20	24.35%	-	2250	73211	¥24.60	¥12.00	26-08-14
Hy3 (high) 推理腾讯混元	36.98	28.15	23.88%	+23.2%	1149	48159	¥5.39	¥4.00	26-07-06
Ling-3.1-flash (xhigh) 推理	41.43	25.10	39.42%	+228%	1033	87880	-	-	26-10-02
GPT-6 Luna (xhigh) 推理 OpenAI	34.47	23.95	30.52%	-	339	27996	¥2.70	¥3.45	26-09-23
MiniMax-M3 推理 MiniMax 海螺	27.85	23.86	14.33%	+33.3%	1150	81187	¥19.10	¥8.40	26-06-01
Seed-2.0-lite 0428 (high) 推理	22.82	16.76	26.56%	-	695	22793	¥2.30	¥3.60	26-05-07
Spark X2.5 推理	24.70	12.71	48.54%	+317%	700	48811	¥4.92	¥3.60	26-09-06
Gemini 3.5 Flash Lite (high) 推理 Google Gemini	18.58	11.92	35.84%	+27.1%	188	15721	¥7.59	¥17.25	26-07-21
GPT-5.5 Instant OpenAI	14.10	9.41	33.26%	-	48	1918	¥11.12	¥207.00	26-05-06
Mistral Large 4 (high) 推理 Mistral AI	10.40	6.41	38.37%	-	561	27865	¥11.25	¥14.42	26-10-06
openPangu-2.0-Pro 推理	9.20	5.98	35.00%	-	752	30015	¥12.19	¥14.50	26-07-31
ERNIE 5.1 推理百度文心	10.79	5.49	49.12%	+2.4%	579	23749	¥11.97	¥18.00	26-05-09
openPangu-2.0-Flash 推理	9.91	5.24	47.12%	-	689	28089	¥1.26	¥1.60	26-06-30
Dots3-Note Preview 推理	9.76	4.91	49.69%	-	233	23477	-	-	26-08-14
Seed-2.1-pro 0915 字节豆包（火山方舟）	8.28	4.91	40.70%	+41.2%	93	1367	¥1.15	¥30.00	26-09-16
Gemini 3.1 Flash Lite Google Gemini	6.21	4.01	35.43%	-	15	677	¥0.20	¥10.35	26-03-03
DeepSeek V4.1 Flash DeepSeek 深度求索	6.71	3.42	49.03%	+56.6%	29	3608	¥0.81	¥8.00	26-09-10
Qwen3.8-27B 阿里云百炼（通义千问）	5.15	2.85	44.66%	-	223	8119	¥2.73	¥12.00	26-08-14
LongCat-2.0 推理美团 LongCat	8.04	2.02	74.88%	-	394	12667	¥2.84	¥8.00	26-06-30