Skip to content

Benchmark

FinBen FPB F1

Scores are percentages. Higher is better.

Models measured
19
Scores recorded
19
Sources
1
IQ weight
0.73

Score distribution

58.5%
Mean
67%
Median
20.9%
Std dev
29%
10th percentile
78%
90th percentile
12.5%
Min
12.5%79%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.73 for chat models, measured over 19 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Llama 3.1 70Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results79%
2GPT-3.5 Turbopublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results78%
3GPT-4 Turbopublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results78%
4gemini-1.0-propublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results77%
5Llama 3.1 8Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results76%
6Qwen2.5 7Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results75.7%
7Qwen 2 (72B)published-demo-leaderboard, FPB, F1· TheFinAI public numeric results75%
8Llama 2 70bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results73%
9codellama-7bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results70.4%
10Mistral (7B) Instruct v0.3published-demo-leaderboard, FPB, F1· TheFinAI public numeric results67%
11Llama 3.2 3Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results63.9%
12code-llama-7b-instructpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results63.6%
13Qwen2 7B Instructpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results52%
14Qwen2.5 7B Instructpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results46.3%
15Llama-2-7b-chat-hfpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results39%
16DeepSeek R1 Distill Qwen 7Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results32%
17Mixtral 8X7b V0.1published-demo-leaderboard, FPB, F1· TheFinAI public numeric results29%
18DeepSeek V3published-demo-leaderboard, FPB, F1· TheFinAI public numeric results24.7%
19DeepSeek R1 Distill Qwen 1.5Bpublished-demo-leaderboard, FPB, F1· TheFinAI public numeric results12.5%

Data credits: The Fin AI. Published FinBen result summaries and Open Greek Financial LLM Leaderboard aggregates. Results licence not separately stated; fractional FinBen success metrics converted to percentages.