Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

DROP leaderboard — the family board.

3 TRACKED VARIANTS · 1 FIRST-CLASS BOARD

One family, 3 tracked prints. Scoring configs and splits of the base board render as sections of this page, each under its own honest label.

● LIVE DATA49 MODELS ON THIS BOARDLEADER DEEPSEEK-V3 AT 91.6

The board — DROPTOP 49 OF 49 MEASURED · BEST SCORE PER MODEL, DEFAULT VARIANT

#ModelVendorScoreVLast seen
01DeepSeek-V3DeepSeek91.62026-08-23
02Llama 3.3 70B InstructMeta90.22026-05-30
03Hunyuan LargeTencent88.92026-05-31
04DeepSeek-V4-Pro-BaseDeepSeek88.72026-06-05
05DeepSeek-V4-Flash-BaseDeepSeek88.62026-06-05
06Claude 3.5 SonnetAnthropic87.12026-08-23
07DeepSeek-V3.2-Exp-BaseDeepSeek86.62026-06-13
08DeepSeek-V3-BaseDeepSeek86.52026-06-05
09Kimi K2 BaseMoonshot86.42026-06-13
10DeepSeek-V3.1-BaseDeepSeek86.32026-06-13
11MiMo V2.5 ProXiaomi86.32026-08-23
12MiMo V2.5 Pro BaseXiaomi86.32026-06-05
13GPT-4 TurboOpenAI862026-08-23
14Qwen 2.5 14BAlibaba85.52026-05-30
15Hy3-preview-BaseTencent85.52026-06-05
16Nova ProAmazon85.42026-09-01
17Llama 3.1 Instruct 405BMeta84.82026-08-23
18MiMo V2 Flash BaseXiaomi84.72026-06-13
19MiMo V2.5 BaseXiaomi83.72026-06-05
20GPT-4oOpenAI83.42026-08-23
21Claude 3.5 HaikuAnthropic83.12026-08-23
22Claude 3 OpusAnthropic83.12026-08-23
23GLM-4.5-BaseZ.ai82.92026-06-05
24GPT-4OpenAI80.92026-08-23
25Nova LiteAmazon80.22026-09-01
26DeepSeek-V2DeepSeek80.12026-05-31
27GPT-4o miniOpenAI79.72026-08-23
28Llama 3.1 70B InstructMeta79.62026-08-23
29Nova MicroAmazon79.32026-09-01
30Claude 3 SonnetAnthropic78.92026-08-23
31Granite 4.1 30B BaseIBM78.62026-06-15
32Claude 3 HaikuAnthropic78.42026-08-23
33Qwen2.5 Instruct 72BAlibaba76.72026-05-30
34Phi 4Microsoft75.52026-08-23
35Gemini 1.5 ProGoogle74.92026-08-23
36Granite 4.1 8B BaseIBM72.42026-06-15
37Llama 3.1 8B InstructMeta71.22026-08-23
38Phi 3 Medium 4K InstructMicrosoft68.32026-05-30
39Granite 4.1 3B BaseIBM66.02026-06-15
40Granite 3.3 8B InstructIBM59.42026-08-23
41Granite 3.1 8B InstructIBM58.62026-06-11
42Granite 3.2 8B InstructIBM58.32026-06-11
43Gemma 3n E2B InstructGoogle53.92026-09-10
44DeepSeek-R1-Distill-Qwen-7BDeepSeek51.82026-06-11
45DeepSeek-R1-Distill-Llama-8BDeepSeek49.72026-06-11
46Granite 3.3 2B InstructIBM44.32026-06-11
47ERNIE 4.5Baidu28.62026-08-23
48Granite 3.2 2B InstructIBM23.82026-06-11
49Granite 3.1 2B InstructIBM21.02026-06-11
◆ IV — INDEPENDENTLY VERIFIED · ◆ AA — AGGREGATOR · ◆ VA — VENDOR · ◇ SA — SOURCE-ATTRIBUTED

Variants of this boardSCORING CONFIGS & SPLITS — NOT CITED STANDALONE, SO THEY LIVE HERE

DROP (3-shot F1)

VARIANT15 models · best score per model
#ModelVendorScoreVLast seen
01DeepSeek-R1DeepSeek92.22026-06-04
02DeepSeek-V3DeepSeek91.62026-06-04
03o1OpenAI90.22026-06-04
04DeepSeek-R1-ZeroDeepSeek89.12026-05-03
05Llama 3.1 Instruct 405BMeta88.72026-05-03
06Claude 3.5 SonnetAnthropic88.32026-06-05
07DeepSeek-V2.5DeepSeek87.82026-05-03
08DeepSeek-R1-Distill-Qwen-14BDeepSeek85.52026-06-05
09o1-miniOpenAI83.92026-06-05
10GPT-4oOpenAI83.72026-06-05
11DeepSeek-V2DeepSeek832026-05-03
12MiMo 7B RLXiaomi78.72026-06-05
13DeepSeek-R1-Distill-Qwen-7BDeepSeek772026-06-05
14Qwen2.5 Instruct 72BAlibaba76.72026-05-03
15QwQ 32B PreviewAlibaba71.22026-06-05

DROP (F1)

VARIANT12 models · best score per model
#ModelVendorScoreVLast seen
01Llama 3.1 Instruct 405BMeta92.52026-06-06
02DeepSeek-V3DeepSeek912026-06-06
03Gemini 2.0 Flash ExpGoogle89.32026-06-06
04GPT-4oOpenAI89.22026-06-06
05Gemini 1.5 ProGoogle89.22026-06-06
06Claude 3.5 SonnetAnthropic88.82026-06-06
07DeepSeek-V4-Pro-BaseDeepSeek88.72026-09-10
08DeepSeek-V4-Flash-BaseDeepSeek88.62026-09-10
09DeepSeek-V3.2-BaseDeepSeek88.22026-06-27
10MiniMax Text 01MiniMax87.82026-06-06
11Qwen2.5 Instruct 72BAlibaba852026-06-06
12DeepSeek-V2DeepSeek80.42026-05-01
THE FAMILY BOARD · EVERY PRINT LABELED · EVERY SCORE SOURCE-ATTRIBUTED ON ITS OWN LIVE BOARDAll leaderboardsThe frontier boardModels by vendor