Skip to content

Benchmark

FinBen MLESG MicroF1

Scores are percentages. Higher is better.

Models measured
18
Scores recorded
18
Sources
1
IQ weight
1.69

Score distribution

28.6%
Mean
33.1%
Median
11.3%
Std dev
17%
10th percentile
40.1%
90th percentile
3%
Min
3%44%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.69 for chat models, measured over 18 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, MLESG, MicroF1· TheFinAI public numeric results44%
2Qwen 2 (72B)published-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results43%
3DeepSeek V3published-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results40.1%
4codellama-7bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results36.1%
5GPT-4 Turbopublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results35%
6gemini-1.0-propublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results34%
7Qwen2 7B Instructpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results34%
8Qwen2.5 7Bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results33.3%
9Mistral (7B) Instruct v0.3published-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results33.1%
10Qwen2.5 7B Instructpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results32.7%
11Llama 2 70bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results31%
12Llama 3.1 8Bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results31%
13GPT-3.5 Turbopublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results25%
14DeepSeek R1 Distill Qwen 7Bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results19.2%
15code-llama-7b-instructpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results17.8%
16Mixtral 8X7b V0.1published-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results17%
17DeepSeek R1 Distill Qwen 1.5Bpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results6.5%
18Llama-2-7b-chat-hfpublished-demo-leaderboard, MLESG, MicroF1· TheFinAI public numeric results3%

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.