Vals CorpFin v2 (CorpFin v2)

We show this table for reference; we do not rank on it.

Data verified August 12, 2026

Vals AI private benchmark for understanding long-context credit agreements.

Evaluation results published by Vals AI. Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.

CorpFin v2 score on CorpFin v2 — August 12, 2026

We mirror the published corpfin v2 score view for CorpFin v2. Claude Opus 5 leads the public snapshot at 73.19%, followed by Claude Fable 5 (71.83%) and Kimi K3 (71.56%). We do not use these results to rank models overall.

1Closed

Claude Opus 5

Anthropic

73.19%

Overall 76.05Context 1M

2Closed

Claude Fable 5

Anthropic

71.83%

Overall 76.28Context 1M+

3Open weights

Kimi K3

Moonshot AI

71.56%

Overall 68.42Context 1.05M

134 modelsKnowledgeCurrentDisplay onlyUpdated August 12, 2026

CorpFin v2 score table (134 models)

CorpFin v2 score results for CorpFin v2
Rank	Model / configuration	Score	Parameters (B)	Open / closed
1	Claude Opus 5Anthropic	73.19%	Not reported	Closed
2	Claude Fable 5Anthropic	71.83%	Not reported	Closed
3	Kimi K3Moonshot AI	71.56%	Not reported	Open
4	Muse Spark 1.1Metaxhigh reasoning	71.29%	Not reported	Closed
5	Muse Spark 1.2Metaxhigh reasoning	70.94%	Not reported	Closed
6	Inkling-SmallThinking Machines Lab0.99 reasoning	69.62%	Not reported	Open
7	InklingThinking Machines Lab0.99 reasoning	68.57%	Not reported	Open
8	Grok 4.3xAI	68.53%	Not reported	Closed
9	GPT-5.5OpenAIxhigh reasoning	68.42%	Not reported	Closed
10	Kimi K2.5 ThinkingMoonshot AI	68.26%	Not reported	Unknown
11	MiniMax M3MiniMax	68.10%	Not reported	Open
12	Qwen3 MaxAlibaba	68.03%	Not reported	Closed
13	Grok 4.5xAIhigh reasoning	67.41%	Not reported	Closed
14	Claude Opus 4.6 (Adaptive)Anthropic	67.02%	Not reported	Closed
15	Claude Sonnet 5Anthropic	66.98%	Not reported	Closed
16	Grok 4 Fast (Reasoning)xAI	66.90%	Not reported	Closed
17	Kimi K2.6Moonshot AI	66.74%	Not reported	Open
18	Claude Opus 4.8Anthropic	66.71%	Not reported	Closed
19	Qwen 3.6 Max (preview)Alibaba	66.47%	Not reported	Closed
20	Gemini 3 Flash PreviewGooglehigh reasoning	66.43%	Not reported	Unknown
21	Grok 4.6xAIhigh reasoning	66.16%	Not reported	Closed
22	GLM-5.2Z.AI	66.12%	Not reported	Open
23	Claude Opus 4.7Anthropic	66.08%	Not reported	Closed
24	Grok 4 0709xAI	66.05%	Not reported	Unknown
25	Grok 4.1 Fast (Reasoning)xAI	65.97%	Not reported	Closed
26	GPT-5.2OpenAIxhigh reasoning	65.89%	Not reported	Closed
27	Qwen3.8 MaxAlibaba	65.85%	Not reported	Open
28	Nemotron 3 Ultra 550b A55bNvidia	65.46%	Not reported	Unknown
29	DeepSeek V4 Pro 0813DeepSeekmax reasoning	65.42%	Not reported	Open
30	GPT-5.6 TerraOpenAIxhigh reasoning	65.31%	Not reported	Closed
31	Claude Sonnet 4.6Anthropic	65.31%	Not reported	Closed
32	Qwen3.5 Plus ThinkingAlibaba	65.31%	Not reported	Unknown
33	GPT-5.4OpenAIxhigh reasoning	65.27%	Not reported	Closed
34	Muse SparkMeta	65.11%	Not reported	Closed
35	Claude Opus 4.5 ThinkingAnthropic	65.07%	Not reported	Closed
36	Gemini 3.5 FlashGooglehigh reasoning	64.69%	Not reported	Closed
37	Gemini 3.1 Pro PreviewGooglehigh reasoning	64.49%	Not reported	Unknown
38	GLM-5.1Z.AI	64.45%	Not reported	Open
39	GPT-5.6 SolOpenAImax reasoning	64.38%	Not reported	Closed
40	GPT-5.6 LunaOpenAImax reasoning	64.22%	Not reported	Closed
41	GPT-5.1OpenAIhigh reasoning	63.83%	Not reported	Closed
42	Qwen3.7 MaxAlibaba	63.71%	Not reported	Closed
43	Grok 4.20 0309 ReasoningxAI	63.67%	Not reported	Unknown
44	Gemini 3 Pro PreviewGooglehigh reasoning	63.67%	Not reported	Unknown
45	Qwen3.5 FlashAlibaba	63.56%	Not reported	Closed
46	Gemini 3.6 FlashGooglehigh reasoning	63.33%	Not reported	Closed
47	GPT-4.1OpenAIhigh reasoning	63.05%	Not reported	Closed
48	GLM 5 ThinkingZhipu AI	62.90%	Not reported	Unknown
49	Qwen3.6-27BAlibaba	62.31%	Not reported	Open
50	Claude Sonnet 4.5 ThinkingAnthropic	61.97%	Not reported	Closed
51	Ling 3.0 Flash 2607Ant	61.93%	Not reported	Unknown
52	Qwen3.6 PlusAlibaba	61.93%	Not reported	Closed
53	DeepSeek V4 Flash 0731DeepSeekhigh reasoning	61.85%	Not reported	Open
54	MiMo-V2.5-ProXiaomi	61.42%	Not reported	Closed
55	DeepSeek V4 Pro 0813DeepSeekmax reasoning	61.38%	Not reported	Open
56	Claude Opus 4.5Anthropic	61.30%	Not reported	Closed
57	Claude Sonnet 4 20250514 ThinkingAnthropic	61.23%	Not reported	Unknown
58	GPT-5.4 nanoOpenAIhigh reasoning	61.19%	Not reported	Closed
59	MiniMax M2.7MiniMax	61.19%	Not reported	Open
60	Grok 3 Mini Fast High ReasoningxAIhigh reasoning	61.11%	Not reported	Unknown
61	GPT-5OpenAIhigh reasoning	61.07%	Not reported	Unknown
62	Mistral Large 2512Mistral AI	61.03%	Not reported	Unknown
63	GLM-4.5Z.AI	60.96%	Not reported	Closed
64	GPT-5.4 miniOpenAIxhigh reasoning	60.92%	Not reported	Closed
65	Claude Sonnet 4.5Anthropic	60.80%	Not reported	Closed
66	Gemini 2.5 ProGoogle	60.80%	Not reported	Closed
67	Gemini 3.5 Flash-LiteGooglehigh reasoning	60.68%	Not reported	Closed
68	Claude Haiku 4.5 ThinkingAnthropic	60.61%	Not reported	Closed
69	Kimi K2 ThinkingMoonshot AI	60.57%	Not reported	Unknown
70	Claude 3.7 Sonnet ThinkingAnthropic	60.41%	Not reported	Closed
71	Claude Haiku 4.5Anthropic	60.30%	Not reported	Closed
72	GPT-5 miniOpenAIhigh reasoning	60.18%	Not reported	Closed
73	MiMo-V2.5Xiaomi	59.91%	Not reported	Closed
74	Gemini 2.5 Pro Exp 03 25Google	59.83%	Not reported	Unknown
75	Gemini 2.5 Flash Preview 09 2025 ThinkingGoogle	59.75%	Not reported	Unknown
76	o3OpenAIhigh reasoning	59.71%	Not reported	Closed
77	Grok 3xAI	59.71%	Not reported	Unknown
78	MiniMax M2.5MiniMax	59.60%	Not reported	Closed
79	Grok 3 Mini Fast Low ReasoningxAIlow reasoning	59.48%	Not reported	Unknown
80	Gemini 3.1 Flash Lite PreviewGooglehigh reasoning	59.36%	Not reported	Unknown
81	o4-miniOpenAIhigh reasoning	58.97%	Not reported	Closed
82	Gemini 2.5 Flash Preview 09 2025Google	58.97%	Not reported	Unknown
83	MiniMax M2.1MiniMax	58.90%	Not reported	Unknown
84	Mistral Medium 3.5Mistral AIhigh reasoning	58.78%	Not reported	Unknown
85	Grok 4 Fast Non ReasoningxAI	58.39%	Not reported	Unknown
86	GPT-OSS 120BOpenAI	58.24%	Not reported	Open
87	Laguna M.1Poolside	58.16%	Not reported	Closed
88	GPT-4.1 miniOpenAIhigh reasoning	57.93%	Not reported	Closed
89	Gemini 2.5 Flash Lite Preview 09 2025 ThinkingGoogle	57.58%	Not reported	Unknown
90	GLM-4.6Z.AI	56.84%	Not reported	Open
91	Laguna XS.2Poolside	56.33%	Not reported	Open
92	Gemini 2.5 Flash Lite Preview 09 2025Google	56.29%	Not reported	Unknown
93	Qwen3 MaxAlibaba	55.94%	Not reported	Closed
94	DeepSeek V3 0324DeepSeek	54.74%	Not reported	Open
95	Claude Sonnet 4Anthropic	54.70%	Not reported	Unknown
96	Trinity-Large-ThinkingArcee AI	54.66%	Not reported	Open
97	Nemotron Lightning 3p5 30b A3bFireworks AI	54.23%	Not reported	Unknown
98	Gemini 2.5 Flash Preview 04 17Google	54.16%	Not reported	Unknown
99	DeepSeek-R1DeepSeek	54.12%	Not reported	Open
100	Claude 3.5 SonnetAnthropic	53.61%	Not reported	Closed
101	GPT-OSS 20BOpenAI	53.15%	Not reported	Open
102	Qwen3 Max PreviewAlibaba	52.95%	Not reported	Unknown
103	Grok 4.1 Fast Non ReasoningxAI	52.49%	Not reported	Unknown
104	DeepSeek V3DeepSeek	52.49%	Not reported	Open
105	DeepSeek V3p1Fireworks AI	51.48%	Not reported	Unknown
106	Grok 2 1212xAI	51.13%	Not reported	Unknown
107	DeepSeek V3p2 ThinkingFireworks AIhigh reasoning	50.97%	Not reported	Unknown
108	Claude 3.5 HaikuAnthropic	50.82%	Not reported	Closed
109	Mistral Medium 2505Mistral AI	50.74%	Not reported	Unknown
110	Moonshotai Kimi K2 InstructTogether AI	50.39%	Not reported	Unknown
111	Llama4 Maverick Instruct BasicFireworks AI	49.73%	Not reported	Unknown
112	DeepSeek V3p2Fireworks AInone reasoning	47.94%	Not reported	Unknown
113	Magistral Medium 2509Mistral AI	47.40%	Not reported	Unknown
114	Meta Llama Llama 4 Scout 17B 16E InstructTogether AI	46.78%	Not reported	Unknown
115	GLM-4.7Z.AI	46.39%	Not reported	Open
116	Command A 03 2025Cohere	45.96%	Not reported	Unknown
117	GPT-4oOpenAIhigh reasoning	45.92%	Not reported	Closed
118	GPT-4o miniOpenAIhigh reasoning	45.45%	Not reported	Closed
119	o3-miniOpenAIhigh reasoning	45.30%	Not reported	Closed
120	Mistral Small 2503Mistral AI	44.17%	Not reported	Unknown
121	Magistral Small 2509Mistral AI	44.02%	Not reported	Unknown
122	Gemini 2.0 Pro Exp 02 05Google	43.44%	Not reported	Unknown
123	GPT-4.1 nanoOpenAIhigh reasoning	42.08%	Not reported	Closed
124	Jamba Large 1.6AI21 Labs	41.53%	Not reported	Unknown
125	Gemini 1.5 Pro 002Google	40.52%	Not reported	Unknown
126	GPT-4oOpenAIhigh reasoning	39.43%	Not reported	Closed
127	Jamba 1.5 LargeAI21 Labs	39.43%	Not reported	Unknown
128	Meta Llama Meta Llama 3.1 70B Instruct TurboTogether AI	38.85%	Not reported	Unknown
129	Gemini 1.5 Flash 002Google	38.19%	Not reported	Unknown
130	Jamba Mini 1.6AI21 Labs	38.03%	Not reported	Unknown
131	Meta Llama Meta Llama 3.1 8B Instruct TurboTogether AI	37.80%	Not reported	Unknown
132	Jamba 1.5 MiniAI21 Labs	33.88%	Not reported	Unknown
133	Gemini 2.0 Flash 001Google	33.72%	Not reported	Unknown
134	Gemini 1.5 Flash 001Google	28.63%	Not reported	Unknown

How CorpFin v2 is shown here

BenchLM mirrors the public Vals AI CorpFin v2 leaderboard captured from https://www.vals.ai/benchmarks/corp_fin_v2 and updated by Vals on August 12, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

CorpFin v2 is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

Snapshot

134 Vals rows4 task viewsprivate datasetTasks: Overall, Exact Pages, Max Fitting Context, Shared Max ContextDisplay only

CorpFin v2 on Vals AI Vals methodology Vals home

The published CorpFin v2 snapshot places Claude Opus 5 first at 73.19%. The third row is 1.63 points behind. The broader top-10 range is 4.94 points, so many of the published results sit in a relatively narrow band.

134 models have been evaluated on CorpFin v2. The benchmark falls in the Knowledge category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. CorpFin v2 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About CorpFin v2

Year

2026

Tasks

Credit-agreement understanding tasks

Format

Accuracy score

Difficulty

Professional finance document reasoning

The Vals CorpFin v2 page reports overall, exact-page, max-fitting-context, and shared-max-context task views. BenchLM keeps it display only.

CorpFin v2 Public benchmark source

Freshness and provenance

Version

CorpFin v2 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does CorpFin v2 measure?

Vals AI private benchmark for understanding long-context credit agreements.

Which model leads the published CorpFin v2 snapshot?

Claude Opus 5 currently leads the published CorpFin v2 snapshot with 73.19% corpfin v2 score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on CorpFin v2?

The August 12, 2026 snapshot contains 134 AI models.

Last updated: August 12, 2026 · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 7,000+ subscribers.

One email each week. Unsubscribe anytime.