VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

DROP

48 models tracked

DROP — Discrete Reasoning Over Paragraphs QA benchmark, unlabeled/default metric

#ModelVendorBest scoreRunsLast seen
1DeepSeek-V3DeepSeek91.612026-08-23
2Llama 3.3 70B InstructMeta90.212026-05-30
3Hunyuan LargeTencent88.912026-05-31
4DeepSeek-V4-Pro-BaseDeepSeek88.732026-06-05
5DeepSeek-V4-Flash-BaseDeepSeek88.632026-06-05
6Claude 3.5 SonnetAnthropic87.112026-08-23
7DeepSeek-V3.2-Exp-BaseDeepSeek86.622026-06-13
8DeepSeek-V3-BaseDeepSeek86.522026-06-05
9Kimi K2 BaseMoonshot86.472026-06-13
10MiMo V2.5 Pro BaseXiaomi86.332026-06-05
11DeepSeek-V3.1-BaseDeepSeek86.322026-06-13
12MiMo V2.5 ProXiaomi86.312026-08-23
13GPT-4 TurboOpenAI8612026-08-23
14Hy3 preview-BaseTencent85.522026-06-05
15Qwen 2.5 14BAlibaba85.512026-05-30
16Nova Pro (Non-Reasoning)Amazon85.412026-08-23
17Llama 3.1 Instruct 405BMeta84.822026-08-23
18MiMo V2 Flash BaseXiaomi84.722026-06-13
19MiMo V2.5 BaseXiaomi83.732026-06-05
20GPT-4oOpenAI83.422026-08-23
21Claude 3.5 HaikuAnthropic83.112026-08-23
22Claude 3 OpusAnthropic83.112026-08-23
23GLM-4.5-BaseZ.ai82.922026-06-05
24GPT4OpenAI80.912026-08-23
25Nova Lite (Non-Reasoning)Amazon80.212026-08-23
26DeepSeek-V2DeepSeek80.112026-05-31
27GPT-4o miniOpenAI79.722026-08-23
28Llama 3.1 70B InstructMeta79.622026-08-23
29Nova Micro (Non-Reasoning)Amazon79.312026-08-23
30Claude 3 SonnetAnthropic78.912026-08-23
31Granite 4.1 30B BaseIBM78.622026-06-15
32Claude 3 HaikuAnthropic78.412026-08-23
33Qwen2.5 Instruct 72BAlibaba76.712026-05-30
34Phi 4Microsoft75.522026-08-23
35Gemini 1.5 ProGoogle74.912026-08-23
36Granite 4.1 8B BaseIBM72.422026-06-15
37Llama 3.1 8B InstructMeta71.222026-08-23
38Phi 3 Medium 4K InstructMicrosoft68.312026-05-30
39Granite 4.1 3B BaseIBM6612026-06-15
40Granite 3.3 8B InstructIBM59.422026-08-23
41Granite 3.1 8B InstructIBM58.612026-06-11
42Granite 3.2 8B InstructIBM58.312026-06-11
43DeepSeek-R1-Distill-Qwen-7BDeepSeek51.812026-06-11
44DeepSeek-R1-Distill-Llama-8BDeepSeek49.712026-06-11
45Granite 3.3 2B InstructIBM44.312026-06-11
46ERNIE 4.5Baidu28.612026-08-23
47Granite 3.2 2B InstructIBM23.812026-06-11
48Granite 3.1 2B InstructIBM2112026-06-11

Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.