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
01Claude Opus 5.5Anthropic86.4BenchAlign推理82.4编程83.5多模态88.8推理82.4编程83.5多模态88.8详情
02GPT-6 AstraOpenAI84.8BenchAlign推理89.6编程76.5多模态—推理89.6编程76.5多模态—详情
03Claude Sonnet 5.5Anthropic83.9BenchAlign推理79.2编程85.0多模态—推理79.2编程85.0多模态—来源
04Gemini 4 ArgonGoogle83.4BenchAlign推理—编程72.8多模态—推理—编程72.8多模态—来源
05Claude Fable 5.1Anthropic81.7BenchAlign推理81.4编程77.8多模态—推理81.4编程77.8多模态—详情
06GPT-6.1 SolOpenAI81.4BenchAlign推理86.8编程71.0多模态—推理86.8编程71.0多模态—来源
07Claude Opus 5Anthropic79.4BenchAlign推理77.3编程71.7多模态88.8推理77.3编程71.7多模态88.8详情
08Claude Fable 5Anthropic78.8BenchAlign推理—编程72.1多模态—推理—编程72.1多模态—详情
09GPT-5.6 SolOpenAI77.9BenchAlign推理72.5编程69.4多模态88.6推理72.5编程69.4多模态88.6详情
10GPT-6 SolOpenAI75.4BenchAlign推理69.4编程66.7多模态—推理69.4编程66.7多模态—详情
11MiMo-V2.6-ProXiaomi74.1BenchAlign推理—编程63.8多模态—推理—编程63.8多模态—来源
12Gemini 3.8 FlashGoogle73.2BenchAlign推理70.6编程62.9多模态83.9推理70.6编程62.9多模态83.9详情
13GPT-5.5 ProOpenAI72.5BenchAlign推理—编程—多模态—推理—编程—多模态—详情
14GPT-5.6 TerraOpenAI72.3BenchAlign推理65.5编程64.2多模态78.2推理65.5编程64.2多模态78.2详情
15Kimi K3Moonshot AI70.7BenchAlign推理65.8编程60.6多模态89.4推理65.8编程60.6多模态89.4详情
16Qwen3.8 MaxAlibaba70.4BenchAlign推理—编程54.5多模态88.4推理—编程54.5多模态88.4详情
17Step 5 PreviewStepFun70.2BenchAlign推理—编程61.7多模态61.2推理—编程61.7多模态61.2来源
18Claude Opus 4.8Anthropic69.1BenchAlign推理59.1编程61.3多模态87.7推理59.1编程61.3多模态87.7详情
19GPT-5.4 ProOpenAI68.9BenchAlign推理—编程—多模态—推理—编程—多模态—详情
20GLM-5.3Z.ai68.6BenchAlign推理77.1编程57.9多模态—推理77.1编程57.9多模态—详情
21Claude Opus 4.7 (Adaptive)Anthropic67.9BenchAlign推理—编程56.4多模态50.1推理—编程56.4多模态50.1详情
22Grok 4.6xAI67.9BenchAlign推理57.7编程60.5多模态—推理57.7编程60.5多模态—详情
23GPT-5.5OpenAI67.7BenchAlign推理66.2编程62.8多模态71.4推理66.2编程62.8多模态71.4详情
24DeepSeek V4.1 FlashDeepSeek67.7BenchAlign推理—编程62.4多模态—推理—编程62.4多模态—来源
25Gemini 3.7 FlashGoogle67.6BenchAlign推理—编程59.2多模态83.6推理—编程59.2多模态83.6详情
26GPT-5.4OpenAI67.0BenchAlign推理60.6编程48.7多模态69.3推理60.6编程48.7多模态69.3详情
27Muse Spark 1.2Meta66.4BenchAlign推理—编程54.4多模态—推理—编程54.4多模态—详情
28GPT-5.2 ProOpenAI66.2BenchAlign推理—编程—多模态—推理—编程—多模态—详情
29GPT-5.6 LunaOpenAI66.0BenchAlign推理54.5编程63.4多模态68.1推理54.5编程63.4多模态68.1详情
30Claude Sonnet 5Anthropic66.0BenchAlign推理—编程58.9多模态78.4推理—编程58.9多模态78.4详情
31Muse Spark 1.1Meta65.9BenchAlign推理—编程54.1多模态—推理—编程54.1多模态—详情
32GPT-6 LunaOpenAI65.6BenchAlign推理55.5编程53.7多模态—推理55.5编程53.7多模态—详情
33MiMo-V2.6-FlashXiaomi65.3BenchAlign推理—编程53.0多模态—推理—编程53.0多模态—来源
34Gemini 3.1 ProGoogle64.6BenchAlign推理—编程41.0多模态80.1推理—编程41.0多模态80.1详情
35Qwen3.8-Flash-NextAlibaba64.2BenchAlign推理—编程53.3多模态84.1推理—编程53.3多模态84.1详情
36Claude Opus 4.7Anthropic64.1BenchAlign推理—编程56.5多模态—推理—编程56.5多模态—详情
37Grok 4.5xAI63.9BenchAlign推理50.4编程57.5多模态—推理50.4编程57.5多模态—详情
38DeepSeek V4 Pro 0813DeepSeek63.9BenchAlign推理—编程48.9多模态—推理—编程48.9多模态—详情
39Gemini 3.6 FlashGoogle63.2BenchAlign推理—编程52.6多模态—推理—编程52.6多模态—详情
40dots3-note PreviewDots Studio63.1BenchAlign推理—编程—多模态—推理—编程—多模态—详情
41Ornith-1.5-397BOrnith AI62.5BenchAlign推理—编程55.1多模态—推理—编程55.1多模态—详情
42Qwen3.7 MaxAlibaba62.5BenchAlign推理—编程45.3多模态—推理—编程45.3多模态—详情
43Gemini 3.5 FlashGoogle62.4BenchAlign推理62.8编程51.5多模态87.8推理62.8编程51.5多模态87.8详情
44Claude Opus 4.6Anthropic62.0BenchAlign推理—编程48.0多模态60.6推理—编程48.0多模态60.6详情
45GLM-5.2Z.ai61.5BenchAlign推理—编程55.2多模态—推理—编程55.2多模态—详情
46Hy4 previewTencent60.9BenchAlign推理—编程58.0多模态—推理—编程58.0多模态—详情
47GPT-5.3 CodexOpenAI60.8BenchAlign推理—编程55.6多模态—推理—编程55.6多模态—详情
48GPT-5.2OpenAI60.3BenchAlign推理—编程38.2多模态67.3推理—编程38.2多模态67.3详情
49Muse SparkMeta60.0BenchAlign推理—编程51.1多模态78.5推理—编程51.1多模态78.5详情
50Gemini 3 ProGoogle59.4BenchAlign推理—编程43.6多模态75.2推理—编程43.6多模态75.2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 —