Skip to content
VECTOR WIREAI INTELLIGENCE
UTC

AGIEval leaderboard — the family board.

4 TRACKED VARIANTS · 1 FIRST-CLASS BOARD

One family, 4 tracked prints. Scoring configs and splits of the base board render as sections of this page, each under its own honest label.

● LIVE DATA4 MODELS ON THIS BOARDLEADER DEEPSEEK-V3 AT 79.6

The board — AGIEvalTOP 4 OF 4 MEASURED · BEST SCORE PER MODEL, DEFAULT VARIANT

#ModelVendorScoreVLast seen
01DeepSeek-V3DeepSeek79.62026-05-01
02Qwen2.5 Instruct 72BAlibaba75.82026-05-01
03Llama 3.1 Instruct 405BMeta60.62026-05-01
04DeepSeek-V2DeepSeek57.52026-05-01
◆ IV — INDEPENDENTLY VERIFIED · ◆ AA — AGGREGATOR · ◆ VA — VENDOR · ◇ SA — SOURCE-ATTRIBUTED

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

AGIEval-En

VARIANT10 models · best score per model
#ModelVendorScoreVLast seen
01Nemotron 3 Ultra 550B A55B BaseNVIDIA78.72026-06-05
02Nemotron 3 Super 120B A12BNVIDIA77.92026-08-11
03Kimi K2 BaseMoonshot72.52026-06-05
04Qwen3.5 35B A3BAlibaba70.52026-08-11
05DeepSeek-V3.2-Exp-BaseDeepSeek70.12026-06-05
06GLM-4.5-BaseZ.ai70.12026-06-05
07Nemotron 3.5 LightningNVIDIA70.02026-08-11
08Mistral Large 3 675B Base 2512Mistral69.32026-06-05
09NVIDIA Nemotron 3 Nano 30B A3BNVIDIA68.52026-08-11
10Gemma 4 26B A4BGoogle55.32026-08-11

AGIEval 5-shot

VARIANT6 models · best score per model
#ModelVendorScoreVLast seen
01Qwen3 14B BaseAlibaba66.12026-06-05
02Ministral 3 14BMistral64.82026-06-12
03Qwen3 8B BaseAlibaba59.62026-06-05
04Ministral 3 8BMistral59.12026-06-05
05Qwen3 4B BaseAlibaba572026-06-05
06Ministral 3 3BMistral51.12026-06-05

Other tracked prints

THIN BOARDSunder 5 published models each — listed, never hidden
PrintLeaderScoreModels
AGIEval (EM)DeepSeek-V4-Pro-Base84.43
THE FAMILY BOARD · EVERY PRINT LABELED · EVERY SCORE SOURCE-ATTRIBUTED ON ITS OWN LIVE BOARDAll leaderboardsThe frontier boardModels by vendor