SuperCLUE-Fin
Models answer Chinese finance questions across 25 tasks in six categories. The evaluation covers financial knowledge and its application in practice.
The authors report graded model comparisons across more than 1,000 questions.
- Made by
- SuperCLUE team
- Released
- April 2024
- Kind
- Finance
- Level
- University
- Scored
- Graded by a model
- Answers
- Open-ended
- Measures
- Capability
- Subject
- Economics and business
- Language
- Chinese
- Interaction
- Multi-turn
- Refresh
- One fixed set
- Run by
- A host
- In views
- Law and finance
- Strongest model tested
- Kimi K3 #25 on the capability index
- Last model added
- July 2026
- Scores refreshed
- November 20th, 2026
Kimi K3 leads with 88.7 among models released by July 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in July 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 Kimi K3 max 88.7
- 2
Qwen3.7-Max 83.6
- 3
Seed2.1 Pro high 83.6
- 4
MiniMax-M3 82.1
- 5
DeepSeek-V4-Pro max 82.0
- 6
Tencent Hy3 high 81.3
- 7 Kimi K2.6 74.7
- 8 GLM-5.2 max 74.6
- 9
LongCat-2.0 73.5
- 10
Step 3.7 Flash high 72.6
- 11 MiMo-V2.5-Pro 72.1
Source: Official SuperCLUE-Fin leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi K3 | 88.7 | max | Jul 2026 | Official leaderboard |
| 2 | | 83.6 | May 2026 | Official leaderboard | |
| 3 | | 83.6 | high | Jun 2026 | Official leaderboard |
| 4 | | 82.1 | Jun 2026 | Official leaderboard | |
| 5 | | 82.0 | max | Apr 2026 | Official leaderboard |
| 6 | | 81.3 | high | Jul 2026 | Official leaderboard |
| 7 | Kimi K2.6 | 74.7 | Apr 2026 | Official leaderboard | |
| 8 | GLM-5.2 | 74.6 | max | Jun 2026 | Official leaderboard |
| 9 | | 73.5 | Jul 2026 | Official leaderboard | |
| 10 | | 72.6 | high | May 2026 | Official leaderboard |
| 11 | MiMo-V2.5-Pro | 72.1 | Apr 2026 | Official leaderboard |
Rankings by part
We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as 金融知识认知, 金融文本理解 and 金融判别与分析. Kimi K3 leads 5 of them, and Seed2.1 Pro leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
金融知识认知
11 models we track have a score here.
- 1 Kimi K3 96.5
- 2 Kimi K2.6 ↑5 93.1
- 3
Tencent Hy3 ↑3 92.8
- 4
Qwen3.7-Max ↓2 92.1
- 5
Seed2.1 Pro ↓2 89.0
- 6
DeepSeek-V4-Pro ↓1 88.4
- 7
MiniMax-M3 ↓3 88.0
- 8
LongCat-2.0 ↑1 85.5
- 9 MiMo-V2.5-Pro ↑2 81.1
- 10
Step 3.7 Flash 79.3
金融文本理解
11 models we track have a score here.
- 1 Kimi K3 94.8
- 2
Seed2.1 Pro ↑1 92.3
- 3
Qwen3.7-Max ↓1 90.9
- 4
Tencent Hy3 ↑2 90.2
- 5
LongCat-2.0 ↑4 88.1
- 6
DeepSeek-V4-Pro ↓1 87.1
- 7
MiniMax-M3 ↓3 86.7
- 8 Kimi K2.6 ↓1 85.7
- 9
Step 3.7 Flash ↑1 85.7
- 10 GLM-5.2 ↓2 84.3
金融判别与分析
11 models we track have a score here.
- 1 Kimi K3 89.8
- 2
Seed2.1 Pro ↑1 87.1
- 3
Tencent Hy3 ↑3 81.8
- 4
Qwen3.7-Max ↓2 81.2
- 5
MiniMax-M3 ↓1 80.5
- 6 GLM-5.2 ↑2 79.9
- 7 Kimi K2.6 78.9
- 8
DeepSeek-V4-Pro ↓3 77.6
- 9
LongCat-2.0 75.9
- 10
Step 3.7 Flash 74.9
金融计算与数值推理
11 models we track have a score here.
- 1
Seed2.1 Pro ↑2 93.6
- 2
DeepSeek-V4-Pro ↑3 93.1
- 3 Kimi K3 ↓2 90.8
- 4
MiniMax-M3 87.0
- 5
Qwen3.7-Max ↓3 86.4
- 6
Tencent Hy3 85.8
- 7 MiMo-V2.5-Pro ↑4 76.6
- 8
LongCat-2.0 ↑1 76.6
- 9
Step 3.7 Flash ↑1 69.1
- 10 GLM-5.2 ↓2 67.6
金融场景应用
11 models we track have a score here.
- 1 Kimi K3 89.3
- 2
Qwen3.7-Max 84.0
- 3
DeepSeek-V4-Pro ↑2 82.6
- 4
MiniMax-M3 82.6
- 5 GLM-5.2 ↑3 79.3
- 6
Tencent Hy3 78.6
- 7
Seed2.1 Pro ↓4 77.3
- 8 Kimi K2.6 ↓1 76.3
- 9 MiMo-V2.5-Pro ↑2 75.9
- 10
Step 3.7 Flash 73.6
金融合规与风控
11 models we track have a score here.
- 1 Kimi K3 82.7
- 2 Kimi K2.6 ↑5 78.8
- 3
Qwen3.7-Max ↓1 78.8
- 4
MiniMax-M3 75.6
- 5
Seed2.1 Pro ↓2 75.0
- 6
Tencent Hy3 75.0
- 7 GLM-5.2 ↑1 72.4
- 8
DeepSeek-V4-Pro ↓3 70.5
- 9
Step 3.7 Flash ↑1 69.9
- 10
LongCat-2.0 ↓1 65.4
4 models have held number one on the official leaderboard since its release in April 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official SuperCLUE-Fin leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official SuperCLUE-Fin leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Law, finance and medicine
- MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy.
Qwen 3.8 Max 80.4%
- HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
- HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
- FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy.
Ling-3.1-flash 81.7%
- KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
- Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark
Catalog record 1 warning 1 pass
- Warning Task access We don’t know whether the tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (11 entries).
- Pass Models scored 11 models we track.
- Pass Newest score The newest is from July 2026, for Kimi K3.
- Note Frontier models None of the top 10 has a score. The strongest tested is Kimi K3 (#25).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on superclueai.com: 11 rows for models we track at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.