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

CRUXEval-O (output prediction)

20 models tracked

CRUXEval-O (output prediction) — CRUXEval output-prediction split; DeepSeek column name for CRUXEval-O

#ModelVendorBest scoreRunsLast seen
1Kimi K2 BaseMoonshot83.542026-06-13
2MiMo V2 Flash BaseXiaomi79.122026-06-13
3DeepSeek-V3.1-BaseDeepSeek76.422026-06-13
4DeepSeek-V3.2-Exp-BaseDeepSeek74.922026-06-13
5DeepSeek-V3-BaseDeepSeek7122026-06-05
6DeepSeek-V3DeepSeek69.812026-05-01
7Hy3 preview-BaseTencent68.422026-06-05
8GLM-4.5-BaseZ.ai67.822026-06-05
9Qwen2.5 7B BaseAlibaba63.412026-05-03
10Stable-DiffCoder-8B-BaseByteDance6012026-06-05
11Qwen2.5 Instruct 72BAlibaba59.912026-05-01
12Llama 3.1 Instruct 405BMeta59.912026-05-01
13Qwen2.5 Coder Instruct 7BAlibaba5612026-06-05
14Granite 4.1 30BIBM55.812026-06-15
15Qwen3 4B Base DenseAlibaba50.712026-05-03
16DeepSeek-V2DeepSeek49.812026-05-01
17Granite 4.1 8BIBM47.612026-06-15
18DeepSeek-Coder-6.7B-BaseDeepSeek4112026-06-05
19Qwen2.5 3B BaseAlibaba26.112026-05-03
20Gemma 3 1B Base DenseGoogle3.712026-05-03

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