Inference
EN
Vals AI

Finance Agent (v2)

Evaluating agents on core financial analyst tasks using the FAB v2 harness

ModelsModels with a result.79
Top resultBest result on this benchmark.65.4%Gemini 4 Argon
Top-3 spreadPoints from first to third.4.1 pts
UpdatedDate the source changed the results.29 Oct 2026

Result and cost

79 of 79 results

Results

79 results
#ModelResultThe score from the source.CostUS dollars to run one task.TimeTime to run one task.
01Gemini 4 Argon65.4% ±0.3$4.3816 min
02Gemini 3.8 Flash61.4% ±0.1$2.003 min
03Muse Spark Open61.3% ±2.1$0.77124 min
04Muse Spark 1.260.6% ±0.3$0.775 min
05Muse Spark 1.3 Max60.0% ±2.1$0.7573 min
06Muse Wonder Full59.1% ±0.2$2.3915 min
07Gemini 3.7 Flash59.0% ±0.3$1.483 min
08Muse Spark 1.358.9% ±0.4$0.7416 min
08Claude Fable 5.158.9% ±2.1$8.3518 min
10Claude Opus 558.6% ±0.1$5.1210 min
10Claude Opus 5.558.6% ±0.2$9.0333 min
12Claude Sonnet 5.558.1% ±0.7$6.3932 min
13Qwen3.8 Max Internal57.9% ±0.1$3.0349 min
13Gemini 3.5 Flash57.9% ±0.2$2.515 min
13GLM-5.3-Flash57.9% ±2.1$0.09515 min
16MiMo-V2.6-Pro57.3% ±0.6$0.19911 min
17Muse Spark 1.157.2% ±0.8$0.7237 min
18Muse Wonder56.6% ±0.4$2.676 min
19Claude Fable 556.3% ±0.8$8.0610 min
19Gemini 3.6 Flash56.3% ±0.2$1.404 min
19MiMo-V2.6-Flash56.3% ±0.5$0.0727 min
22GLM-5.355.8% ±2.1$1.0716 min
23Hy4 preview55.1% ±0.3$0.5924 min
24GPT-5.6 Luna55.0% ±0.3$0.27913 min
25Ling 3.0 Flash Af Rc354.9% ±0.7$0.0456 min
26Mistral Large 454.7% ±0.6$0.61526 min
27GPT-5.6 Terra54.4% ±2.1$3.6324 min
28Claude Haiku 5.554.1% ±2.1$1.6431 min
29Claude Opus 4.853.9% ±0.2$4.229 min
29Claude Sonnet 553.9% ±0.5$0.75113 min
31GPT-5.6 Sol53.8% ±0.8$1.2519 min
32Grok 4.653.7% ±0.7$1.6617 min
33GPT-6 Astra53.5% ±2.1$6.8212 min
33DeepSeek V4.1 Flash 101753.5% ±0.4$0.2076 min
35Kimi K353.1% ±0.4$1.915 min
36Grok 4.752.3% ±0.4$2.7216 min
37GPT-6.1 Sol52.0% ±0.3$1.6215 min
38Ember-151.9% ±2.1$1.9513 min
39GPT-5.551.8% ±0.6$4.1511 min
40Claude Opus 4.751.5% ±0.5$4.036 min
41Claude Sonnet 4.651.0% ±0.3$2.4112 min
42Step 5 Preview50.7% ±0.3$0.61912 min
43Qwen3.8 Max50.6% ±0.6$1.2420 min
44DeepSeek V4 Pro 081350.4% ±0.2$0.8816 min
45GPT-6 Luna49.9% ±0.2$0.12415 min
46GLM-5.249.7% ±0.9$0.7148 min
47DeepSeek V4 Flash 073149.5% ±0.5$0.2768 min
48GPT-6 Sol49.0% ±0.6$2.129 min
49Qwen3.8-27B48.6% ±2.1$0.74910 min
50Grok 4.548.3% ±0.5$1.165 min
50MiniMax M348.3% ±0.4$0.3238 min
52Qwen3.7 Max47.8% ±0.7$0.55311 min
53Gemini 3.5 Flash-Lite47.4% ±0.5$0.392 min
54Inkling46.6% ±0.9$0.99618 min
55GPT-5.4 mini45.4% ±0.4$1.2036 min
56Kimi K2.644.9% ±0.7$1.1427 min
57GLM-5.144.8% ±0.6$0.61815 min
58DeepSeek V4 Pro 042344.1% ±0.7$0.87925 min
59Gemini 3.1 Pro Preview43.0% ±1.2$1.593 min
60Gemini 3 Flash Preview42.6% ±0.4$0.5714 min
61MiMo-V2.5-Pro41.5% ±1.2$0.2068 min
62Inkling-Small41.3% ±0.8$0.2317 min
63Qwen3.6 Plus40.8% ±0.1$0.81410 min
64Qwen3.7 Plus38.2% ±1.0$0.3589 min
64GPT-5.4 nano38.2% ±1.2$0.1596 min
66Grok 4.337.7% ±0.4$0.84810 min
67MiMo-V2.536.7% ±0.1$0.0876 min
68Kimi K2.5 Thinking35.8% ±2.0$0.29413 min
69Mistral Medium 3.532.1% ±1.0$1.9113 min
70Claude Haiku 4.5 Thinking31.0% ±0.6$0.6023 min
71Ling 3.0 Flash30.3% ±0.3$0.053 min
72Gemini 3.1 Flash Lite Preview30.0% ±0.6$0.14255 s
73Grok 4.20 (Reasoning)28.5% ±0.3$0.7926 min
74MiniMax M2.727.9% ±0.5$0.18313 min
75Laguna M.125.0% ±1.3$0.81820 min
76Mercury 2.518.9% ±0.2$0.1062 min
77Nemotron Lightning 3p5 30B A3B18.5% ±0.9$0.0299 min
78Laguna XS.215.6% ±0.3$0.1929 min
79Command A+9.0% ±0.4$1.639 min