Skip to content

Benchmark

FinBen FiQA-SA F1

Scores are percentages. Higher is better.

Models measured
19
Scores recorded
19
Sources
1
IQ weight
0.30

Score distribution

58.5%
Mean
65.7%
Median
20.9%
Std dev
22.6%
10th percentile
80%
90th percentile
16%
Min
16%83%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.30 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 2 70bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results83%
2gemini-1.0-propublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results81%
3GPT-4 Turbopublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results80%
4Llama-2-7b-chat-hfpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results76%
5Llama 3.1 8Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results75%
6Llama 3.2 3Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results74.4%
7Llama 3.1 70Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results74%
8DeepSeek V3published-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results66.4%
9Mistral (7B) Instruct v0.3published-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results66.2%
10code-llama-7b-instructpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results65.7%
11Qwen 2 (72B)published-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results63%
12GPT-3.5 Turbopublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results60%
13Qwen2 7B Instructpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results57%
14DeepSeek R1 Distill Qwen 7Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results54.6%
15Qwen2.5 7Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results42.7%
16DeepSeek R1 Distill Qwen 1.5Bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results34.8%
17codellama-7bpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results22.6%
18Qwen2.5 7B Instructpublished-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results19.1%
19Mixtral 8X7b V0.1published-demo-leaderboard, FiQA-SA, F1· TheFinAI public numeric results16%

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.