BenchLM 语言模型排行榜

按 BenchAlign 综合分比较语言模型,查看推理、编程与多模态表现。

评测来源 BenchLM采集 / 更新 查看原榜评分方法

BenchAlign

208 个模型
BenchLM · BenchAlign
排名模型 / 开发机构BenchAlign推理编程多模态详情
01GPT-6 AstraOpenAI88.8BenchAlign推理89.6编程74.0多模态—推理89.6编程74.0多模态—详情
02Claude Opus 5.5Anthropic87.8BenchAlign推理82.4编程83.3多模态88.8推理82.4编程83.3多模态88.8详情
03Claude Sonnet 5.5Anthropic83.4BenchAlign推理79.2编程85.6多模态—推理79.2编程85.6多模态—来源
04Claude Fable 5.1Anthropic82.8BenchAlign推理81.4编程79.7多模态—推理81.4编程79.7多模态—详情
05Claude Opus 5Anthropic80.0BenchAlign推理77.3编程72.0多模态88.8推理77.3编程72.0多模态88.8详情
06Claude Fable 5Anthropic79.5BenchAlign推理—编程73.0多模态—推理—编程73.0多模态—详情
07GPT-6 SolOpenAI79.2BenchAlign推理79.6编程61.8多模态—推理79.6编程61.8多模态—详情
08GPT-5.6 SolOpenAI79.0BenchAlign推理72.5编程70.5多模态88.6推理72.5编程70.5多模态88.6详情
09GPT-6.1 SolOpenAI77.6BenchAlign推理79.4编程59.0多模态—推理79.4编程59.0多模态—来源
10Gemini 4 ArgonGoogle77.2BenchAlign推理—编程74.0多模态—推理—编程74.0多模态—来源
11MiMo-V2.6-ProXiaomi75.5BenchAlign推理—编程68.0多模态—推理—编程68.0多模态—来源
12GPT-5.5 ProOpenAI74.6BenchAlign推理—编程—多模态—推理—编程—多模态—详情
13Gemini 3.8 FlashGoogle74.0BenchAlign推理70.6编程63.6多模态83.9推理70.6编程63.6多模态83.9详情
14GPT-5.6 TerraOpenAI73.2BenchAlign推理65.5编程64.3多模态78.2推理65.5编程64.3多模态78.2详情
15Kimi K3Moonshot AI72.1BenchAlign推理65.8编程61.4多模态89.4推理65.8编程61.4多模态89.4详情
16Qwen3.8 MaxAlibaba72.1BenchAlign推理—编程55.4多模态88.4推理—编程55.4多模态88.4详情
17Claude Opus 4.8Anthropic70.8BenchAlign推理59.1编程62.2多模态87.7推理59.1编程62.2多模态87.7详情
18GPT-5.4 ProOpenAI70.7BenchAlign推理—编程—多模态—推理—编程—多模态—详情
19Step 5 PreviewStepFun70.4BenchAlign推理—编程61.4多模态61.2推理—编程61.4多模态61.2来源
20Claude Opus 4.7 (Adaptive)Anthropic69.7BenchAlign推理—编程57.5多模态50.1推理—编程57.5多模态50.1详情
21GPT-5.5OpenAI69.4BenchAlign推理66.2编程62.7多模态71.4推理66.2编程62.7多模态71.4详情
22Grok 4.6xAI69.4BenchAlign推理57.7编程61.4多模态—推理57.7编程61.4多模态—详情
23GPT-5.4OpenAI68.9BenchAlign推理60.6编程49.6多模态69.3推理60.6编程49.6多模态69.3详情
24GPT-5.2 ProOpenAI68.2BenchAlign推理—编程—多模态—推理—编程—多模态—详情
25Gemini 3.7 FlashGoogle68.0BenchAlign推理—编程59.3多模态83.6推理—编程59.3多模态83.6详情
26Claude Sonnet 5Anthropic67.4BenchAlign推理—编程59.8多模态78.4推理—编程59.8多模态78.4详情
27Muse Spark 1.1Meta66.6BenchAlign推理—编程55.3多模态—推理—编程55.3多模态—详情
28MiMo-V2.6-FlashXiaomi66.4BenchAlign推理—编程53.1多模态—推理—编程53.1多模态—来源
29GPT-5.6 LunaOpenAI66.2BenchAlign推理54.5编程63.3多模态68.1推理54.5编程63.3多模态68.1详情
30Claude Opus 4.7Anthropic65.8BenchAlign推理—编程57.9多模态—推理—编程57.9多模态—详情
31Muse Spark 1.2Meta65.7BenchAlign推理—编程55.0多模态—推理—编程55.0多模态—详情
32GLM-5.3Z.ai65.7BenchAlign推理77.1编程56.6多模态—推理77.1编程56.6多模态—详情
33Grok 4.5xAI65.7BenchAlign推理50.4编程58.1多模态—推理50.4编程58.1多模态—详情
34Gemini 3.6 FlashGoogle65.2BenchAlign推理—编程53.6多模态—推理—编程53.6多模态—详情
35Gemini 3.1 ProGoogle65.1BenchAlign推理—编程42.3多模态80.1推理—编程42.3多模态80.1详情
36GPT-6 LunaOpenAI65.0BenchAlign推理55.6编程50.2多模态—推理55.6编程50.2多模态—详情
37DeepSeek V4 Pro 0813DeepSeek65.0BenchAlign推理—编程49.2多模态—推理—编程49.2多模态—详情
38DeepSeek V4.1 FlashDeepSeek64.7BenchAlign推理—编程58.8多模态—推理—编程58.8多模态—来源
39Qwen3.8-Flash-NextAlibaba64.5BenchAlign推理—编程55.3多模态84.1推理—编程55.3多模态84.1详情
40Claude Opus 4.6Anthropic64.3BenchAlign推理—编程49.5多模态60.6推理—编程49.5多模态60.6详情
41Gemini 3.5 FlashGoogle64.0BenchAlign推理62.8编程52.3多模态87.8推理62.8编程52.3多模态87.8详情
42dots3-note PreviewDots Studio63.6BenchAlign推理—编程—多模态—推理—编程—多模态—详情
43Qwen3.7 MaxAlibaba63.5BenchAlign推理—编程45.4多模态—推理—编程45.4多模态—详情
44GLM-5.2Z.ai63.2BenchAlign推理—编程56.2多模态—推理—编程56.2多模态—详情
45Ornith-1.5-397BOrnith AI62.7BenchAlign推理—编程54.7多模态—推理—编程54.7多模态—详情
46GPT-5.3 CodexOpenAI62.5BenchAlign推理—编程56.0多模态—推理—编程56.0多模态—详情
47GPT-5.2OpenAI61.9BenchAlign推理—编程39.6多模态67.3推理—编程39.6多模态67.3详情
48Gemini 3 ProGoogle61.6BenchAlign推理—编程45.4多模态75.2推理—编程45.4多模态75.2详情
49Muse SparkMeta61.3BenchAlign推理—编程53.1多模态78.5推理—编程53.1多模态78.5详情
50Hy4 previewTencent61.1BenchAlign推理—编程60.6多模态—推理—编程60.6多模态—详情
显示更多已显示 50 / 208

保留来源评分口径,请在同一来源、同一任务内比较。

未公布的数据记为 —