Excel Modeling Benchmark
Results
26 of 41 tracked models have a result.
- 1
Anthropic · Max effort
76.7% - 2
Anthropic · Max effort
75.9% - 3
Anthropic · Max effort
75.7% - 4Gemini 4 ArgonLimited
Google DeepMind · High effort
75.2% - 5
Google DeepMind · High effort
72.2% - 6
OpenAI · Max effort
71.7% - 7
OpenAI · Max effort
70.8% - 8
OpenAI · Max effort
68.5% - 9
Meta · Max effort
67.4% - 10
SpaceXAI · xHigh effort
67.0% - 11Kimi K3Open
Moonshot AI · Max effort
66.7% - 12
OpenAI · Max effort
66.2% - 13MiMo-V2.6-FlashOpen
Xiaomi
65.5% - 14MiMo-V2.6-ProOpen
Xiaomi
62.9% - 15
Alibaba
60.1% - 16
Anthropic · Max effort
59.0% - 17Hy4 previewOpen
Tencent
58.1% - 18
DeepSeek · High effort
57.2% - 19
StepFun
57.1% - 20GLM-5.3Open
Z.ai · Max effort
56.3% - 21
Mistral AI · High effort
56.2% - 22DeepSeek V4 ProOpen
DeepSeek · Max effort
52.8% - 23
Google DeepMind · High effort
52.6% - 24MiniMax-M3Open
MiniMax
47.8% - 25InklingOpen
Thinking Machines Lab
38.4% - 26Inkling-SmallOpen
Thinking Machines Lab
31.8%
15 tracked models without a result
Command A+, Gemini 3.5 Flash-Lite, Gemma 4 31B, GLM-5.3-Flash, gpt-oss-120b, Hy3, Kimi K2.7 Code, MAI-Thinking-1, Mistral Medium 3.5, Mistral Small 4, Muse Glimmer 30B, Nemotron 3 Ultra, Nova 2 Lite, Qwen3.8-27B, Qwen3.8-Flash.
Missing means the evaluator has not published a run, not that the model would score poorly.