Vals AI

CorpFin v2

A private benchmark evaluating understanding of long-context credit agreements

Not in index
Finance
Source ↗

Models126

Top result73.2%Claude Opus 5

Top-3 spread1.9 pts

Updated12 Aug 2026

Result and cost

Results

127 results

#ModelResultCostTime

Claude Opus 573.2%$0.8629 s

Claude Fable 571.8%$1.7453 s

Muse Spark 1.171.3%$0.10928 s

Muse Spark 1.270.9%$0.10830 s

05TMInkling-Small69.6%$0.02545 s

06TMInkling68.6%$0.08429 s

Grok 4.368.5%$0.13436 s

GPT-5.568.4%$0.43223 s

09MAKimi K2.5 Thinking68.3%$0.05281 s

MiniMax M368.1%$0.0524 s

Grok 4.567.4%$0.2212 min

Claude Opus 4.6 (Adaptive)67.0%$0.48721 s

Claude Sonnet 567.0%$0.51619 s

Grok 4 Fast (Reasoning)66.9%$0.0412 s

Claude Opus 4.866.7%$0.85620 s

Qwen 3.6 Max (preview)66.5%$0.15647 s

Gemini 3 Flash Preview66.4%$0.04711 s

Grok 4.666.2%$0.16624 s

GLM-5.266.1%$0.12225 s

Claude Opus 4.766.1%$0.81618 s

Grok 466.0%$0.19130 s

Grok 4.1 Fast (Reasoning)66.0%$0.01628 s

GPT-5.265.9%$0.15426 s

Qwen3.8 Max65.9%$0.17947 s

GPT-5.6 Terra65.3%$0.1717.7 s

Claude Sonnet 4.665.3%$0.30718 s

Qwen3.5 Plus Thinking65.3%$0.0583 min

GPT-5.465.3%$0.23582 s

Muse Spark65.1%$0.01240 s

Claude Opus 4.5 Thinking65.1%$0.17318 s

Gemini 3.5 Flash64.7%$0.148.7 s

Gemini 3.1 Pro Preview64.5%$0.18225 s

GLM-5.164.5%$0.07968 s

GPT-5.6 Sol64.4%$0.48516 s

GPT-5.6 Luna64.2%$0.01917 s

GPT-5.163.8%$0.09235 s

Qwen3.7 Max63.7%$0.2220 s

Gemini 3 Pro63.7%$0.17920 s

Grok 4.20 (Reasoning)63.7%$0.1639.7 s

Qwen3.5 Flash63.6%<$0.0135 s

DeepSeek V4 Pro 042363.4%<$0.0133 s

Gemini 3.6 Flash63.3%$0.1376.3 s

GPT-4.163.1%$0.13210 s

GLM-5 Thinking62.9%$0.06869 s

Qwen3.6-27B62.3%$0.05463 s

Qwen3 Maxqwen3-max-2026-01-2362.0%$0.2612 min

Claude Sonnet 4.5 Thinking62.0%$0.25521 s

Qwen3.6 Plus61.9%$0.04736 s

48INLing 3.0 Flash61.9%<$0.0112 s

DeepSeek V4 Flash 042361.8%$0.01216 s

MiMo-V2.5-Pro61.4%$0.03719 s

Claude Opus 4.561.3%$0.159.4 s

Claude Sonnet 4 Thinking61.2%$0.2373 min

MiniMax M2.761.2%$0.02311 s

GPT-5.4 nano61.2%$0.0169.4 s

Grok 3 Mini Fast High Reasoning61.1%$0.01714 s

GPT-561.1%$0.0863 min

Mistral Large61.0%$0.02449 s

GLM-4.561.0%$0.0382 min

GPT-5.4 mini60.9%$0.07652 s

Claude Sonnet 4.560.8%$0.23713 s

Gemini 2.5 Pro60.8%$0.09527 s

Gemini 3.5 Flash-Lite60.7%$0.0293.8 s

Claude Haiku 4.5 Thinking60.6%$0.08914 s

65MAKimi K2 Thinking60.6%$0.0382 min

Claude 3.7 Sonnet Thinking60.4%$0.23222 s

Claude Haiku 4.560.3%$0.0768.0 s

GPT-5 mini60.2%$0.02174 s

MiMo-V2.559.9%$0.01218 s

Gemini 2.5 Pro Exp 03.2559.8%$0.08616 s

Gemini 2.5 Flash Preview Thinking59.8%$0.02652 s

o359.7%$0.13319 s

Grok 3 [Beta]59.7%$0.1724 s

MiniMax M2.559.6%$0.02417 s

Grok 3 Mini Fast Low Reasoning59.5%$0.01711 s

Gemini 3.1 Flash Lite Preview59.4%$0.0234.0 s

o4-mini (high)59.0%$0.07418 s

Gemini 2.5 Flash (Preview)59.0%$0.02652 s

MiniMax M2.158.9%$0.02222 s

Mistral Medium 3.558.8%$0.14644 s

Grok 4 Fast Non Reasoning58.4%$0.049.4 s

GPT-OSS 120B58.2%<$0.0143 s

83POLaguna M.158.2%—4 min

GPT-4.1 mini57.9%$0.0266.4 s

Gemini 2.5 Flash Lite Preview Thinking57.6%<$0.0154 s

GLM-4.656.8%$0.04578 s

87POLaguna XS.256.3%—3 min

Gemini 2.5 Flash Lite Preview56.3%<$0.0155 s

Claude Sonnet 454.7%$0.2273 min

90AATrinity-Large-Thinking54.7%$0.01912 s

Nemotron Lightning 3p5 30B A3B54.2%<$0.0139 s

Gemini 2.5 Flash Preview 04.1754.2%$0.0219.0 s

DeepSeek-R154.1%$0.08747 s

Claude 3.5 Sonnet53.6%$0.2380.0 s

DeepSeek V353.6%$0.05237 s

GPT-OSS 20B53.1%<$0.0134 s

Qwen3 Maxqwen3-max-preview53.0%$0.0974 min

Grok 4.1 Fast Non Reasoning52.5%$0.01613 s

DeepSeek V3.151.5%$0.03763 s

100

Grok 251.1%$0.11587 s

101

DeepSeek V3.2 (Thinking)51.0%$0.0382 min

102

Claude 3.5 Haiku50.8%$0.0630.0 s

103

Mistral Medium50.7%$0.02430 s

104MAKimi K2 Instruct50.4%$0.05941 s

105

Llama4 Maverick Instruct Basic49.7%$0.0144.7 s

106

DeepSeek V3.247.9%$0.0372 min

107

Magistral Medium47.4%$0.12555 s

108

Llama 4 Scout 17B 16E Instruct46.8%$0.0129.7 s

109

GLM-4.746.4%$0.03966 s

110COCommand A46.0%$0.20614 s

111

GPT-4o mini45.5%<$0.010.0 s

112

o3-mini45.3%$0.07831 s

113

Mistral Small44.2%<$0.0111 s

114

Magistral Small44.0%$0.0314 s

115

Gemini 2.0 Pro Exp 02.0543.4%$0.10519 s

116

GPT-4o42.7%—6.5 s

117

GPT-4.1 nano42.1%<$0.015.0 s

118ALJamba Large 1.641.5%$0.1527 s

119

Gemini 1.5 Pro 00240.5%$0.13138 s

120ALJamba 1.5 Large39.4%$0.14911 s

121

Meta Llama 3.1 70B Instruct Turbo38.9%$0.0530.1 s

122

Gemini 1.5 Flash 00238.2%<$0.0128 s

123ALJamba Mini 1.638.0%$0.0154.2 s

124

Meta Llama 3.1 8B Instruct Turbo37.8%$0.0110.0 s

125ALJamba 1.5 Mini33.9%$0.0152.3 s

126

Gemini 2.0 Flash 00133.7%$0.0130 s

127

Gemini 1.5 Flash 00128.6%<$0.011.2 s