Inference
EN
Vals AI

EMB

Evaluating agents on Excel-based financial modeling tasks

ModelsModels with a result.75
Top resultBest result on this benchmark.76.7%Claude Fable 5.1
Top-3 spreadPoints from first to third.1.0 pts
UpdatedDate the source changed the results.20 Oct 2026

Result and cost

75 of 75 results

Results

75 results
#ModelResultThe score from the source.CostUS dollars to run one task.TimeTime to run one task.
01Claude Fable 5.176.7% ±2.1$15.9330 min
02Claude Opus 5.575.9% ±2.4$9.3335 min
03Claude Sonnet 5.575.7% ±2.4$6.5337 min
04Gemini 4 Argon75.2% ±2.5$11.8038 min
05Claude Fable 573.7% ±2.5$12.3527 min
06Claude Opus 573.6% ±2.2$8.0226 min
07GPT-5.6 Sol72.3% ±2.4$6.0110 min
08Gemini 3.8 Flash72.2% ±2.4$8.2413 min
09Qwen3.8 Max Internal72.0% ±2.4$4.4353 min
10Muse Wonder Full71.9% ±2.7$4.5828 min
11GPT-6 Astra71.7% ±2.5$5.8313 min
12GPT-6 Sol71.5% ±2.4$1.2810 min
13Gemini 3.7 Flash71.3% ±2.3$5.6911 min
14GPT-6.1 Sol70.8% ±2.7$1.1418 min
15Claude Opus 4.869.4% ±2.6$12.0644 min
16GPT-6 Luna68.5% ±2.8$0.14217 min
17Muse Spark 1.3 Max67.4% ±3.1$2.5514 min
18GPT-5.6 Luna67.1% ±2.9$0.36711 min
18Muse Wonder67.1% ±2.9$5.6017 min
20Grok 4.767.0% ±3.0$6.4835 min
21Kimi K366.7% ±2.9$3.0712 min
22Claude Sonnet 566.3% ±3.0$10.2950 min
23GPT-5.6 Terra66.2% ±3.1$2.1914 min
24MiMo-V2.6-Flash65.5% ±2.7$0.10925 min
25Gemini 3.6 Flash65.4% ±2.6$3.8912 min
26GPT-5.564.5% ±2.9$3.2715 min
27Claude Opus 4.763.7% ±3.1$12.7639 min
28Gemini 3.5 Flash63.5% ±2.8$5.1719 min
29Ember-163.4% ±2.9$2.4631 min
30MiMo-V2.6-Pro62.9% ±3.1$0.32955 min
31Grok 4.662.7% ±3.1$3.0653 min
31Muse Spark 1.362.7% ±3.0$3.5523 min
33GLM-5.261.5% ±2.9$4.4759 min
34Claude Sonnet 4.660.2% ±3.3$7.2354 min
35Qwen3.8 Max60.1% ±3.4$2.6658 min
36Qwen3.8-27B59.7% ±3.0$4.5623 min
37Claude Haiku 5.559.0% ±4.2$1.1518 min
38Hy4 preview58.1% ±3.1$0.78731 min
39Kimi K2.657.9% ±2.7$2.2040 min
40DeepSeek V4.1 Flash 101757.2% ±3.2$0.23413 min
41Step 5 Preview57.1% ±3.2$1.3344 min
42Qwen3.7 Max57.0% ±2.9$2.8633 min
42Muse Spark 1.257.0% ±3.1$2.3117 min
42DeepSeek V4 Flash 073157.0% ±3.4$0.38426 min
45Muse Spark 1.156.4% ±3.2$1.1115 min
46GLM-5.356.3% ±3.3$3.7936 min
47Mistral Large 456.2% ±3.2$4.8482 min
48GLM-5.3-Flash55.9% ±3.6$0.71375 min
49MiMo-V2.5-Pro55.2% ±3.1$0.22129 min
50MiMo-V2.555.1% ±3.1$0.08331 min
51Muse Spark Open53.7% ±3.2$5.3585 min
52Grok 4.552.9% ±3.1$1.5518 min
53DeepSeek V4 Pro 081352.8% ±3.1$1.0837 min
54Gemini 3.1 Pro Preview52.6% ±3.0$3.7712 min
55DeepSeek V4 Pro 042351.6% ±3.1$0.79714 min
56Qwen3.7 Plus49.3% ±3.2$0.59837 min
57MiniMax M347.8% ±2.9$2.0932 min
58GPT-5.4 mini45.4% ±3.6$2.1543 min
59GPT-5.4 nano44.8% ±3.1$0.57329 min
60Gemini 3.5 Flash-Lite43.2% ±3.1$1.238 min
61Inkling38.4% ±3.3$1.6726 min
62Ling 3.0 Flash Af Rc336.8% ±3.3$0.1837 min
63Qwen3.6 Plus32.9% ±2.8$2.8731 min
64Inkling-Small31.8% ±3.0$0.24920 min
65Kimi K2.5 Thinking28.5% ±2.2$0.78763 min
65MiniMax M2.728.5% ±2.4$1.6620 min
67Gemini 3 Flash Preview26.0% ±2.9$0.4535 min
68Ling 3.0 Flash25.0% ±2.5$0.1812 min
69Claude Haiku 4.5 Thinking22.7% ±2.4$1.0312 min
70Grok 4.318.4% ±1.7$0.6345 min
71Nemotron Lightning 3p5 30B A3B17.0% ±1.5$0.18239 min
72Grok 4.20 (Reasoning)11.7% ±1.6$0.476 min
73Mistral Medium 3.511.2% ±1.6$10.6051 min
74Mercury 2.58.8% ±1.3$0.4185 min
75Gemini 3.1 Flash Lite Preview8.6% ±1.1$0.06967 s