FinanceBenchmark: Financial AI Leaderboard 2026

FinanceBenchmark provides independent rankings and a composite score for leading Artificial Intelligence models, including Anthropic Claude, OpenAI GPT, and Google Gemini, on complex financial reasoning and forensic accounting tasks. Enable JavaScript to view the interactive, continuously updated leaderboard.

2026-09-10MODELDeepSeek-V4.1-Flash released with extreme agent memory efficiency and evaluated across VALS • 2026-09-03MODELGPT-6 Astra released by OpenAI - 71.7% on EMB, 63.3% on Tax Agent Bench • 2026-09-02MODELGoogle Gemini 3.8 Flash takes #1 rank on Finance Agent (v2) with 61.4% and 72.2% on EMB • 2026-09-02MODELMeta releases Muse Spark 1.3 & 1.3 Max - 59.0%-60.0% on Finance Agent (v2) • 2026-09-01MODELClaude Fable 5.1 leads Tax Agent Bench (77.6%) and EMB (76.7%) with 75% cheaper cache reads • 2026-09-11BENCHTax Agent Bench and FAB v2 benchmarks updated with latest frontier model rankings • 2026-09-15UPDATEPlatform updated: 43 models tracked across 41 independent benchmarks • 2026-09-10MODELDeepSeek-V4.1-Flash released with extreme agent memory efficiency and evaluated across VALS • 2026-09-03MODELGPT-6 Astra released by OpenAI - 71.7% on EMB, 63.3% on Tax Agent Bench • 2026-09-02MODELGoogle Gemini 3.8 Flash takes #1 rank on Finance Agent (v2) with 61.4% and 72.2% on EMB • 2026-09-02MODELMeta releases Muse Spark 1.3 & 1.3 Max - 59.0%-60.0% on Finance Agent (v2) • 2026-09-01MODELClaude Fable 5.1 leads Tax Agent Bench (77.6%) and EMB (76.7%) with 75% cheaper cache reads • 2026-09-11BENCHTax Agent Bench and FAB v2 benchmarks updated with latest frontier model rankings • 2026-09-15UPDATEPlatform updated: 43 models tracked across 41 independent benchmarks •

Tier 3documentQAforensic

LegalBench

Collaboratively built benchmark for measuring legal reasoning in large language models.

Variable tasks

21 models evaluated

Updated 2026

arXiv:2308.11462

View Source →Submit your score Methodology

88.6%

Top Score

Top model

Claude Fable 5

Model Rankings

financebenchmark.ai · LegalBench

86.0%

🥇

AnthropicClaude Opus 587.0%

🥈

AnthropicClaude Sonnet 583.9%

🥉

OpenAIGPT-5.6 Sol87.0%

4

DeepSeekDeepSeek V4 FlashOPEN77.7%

5

xAIGrok 4.686.3%

6

Moonshot AIKimi K3OPEN86.0%

7

AnthropicClaude Fable 588.6%

8

DeepSeekDeepSeek V4 Pro (0813)OPEN82.4%

9

GoogleGemini 3.5 Flash Lite84.1%

10

GoogleGemini 3.6 Flash86.5%

11

Zhipu AIGLM 5.2OPEN84.1%

12

OpenAIGPT-5.6 Luna84.0%

13

OpenAIGPT-5.6 Terra85.1%

14

xAIGrok 4.586.0%

15

Thinking MachinesInklingOPEN83.0%

16

Thinking MachinesInkling SmallOPEN83.0%

17

Ant GroupLing 3.0 FlashOPEN79.7%

18

MetaMuse Spark 1.185.0%

19

MetaMuse Spark 1.285.3%

20

NVIDIANemotron 3 UltraOPEN82.1%

21

AlibabaQwen 3.8 MaxOPEN83.6%

Full Results Table

#	Model	Overall	Source	Date
🥇	Claude Fable 5 Anthropic	88.6	arXiv:2308.11462	2026-06
🥈	Claude Opus 5 Anthropic	87.0	arXiv:2308.11462	2026-07
🥉	GPT-5.6 Sol OpenAI	87.0	arXiv:2308.11462	2026-07
4	Gemini 3.6 Flash Google	86.5	arXiv:2308.11462	2026-07
5	Grok 4.6 xAI	86.3	arXiv:2308.11462	2026-08
6	Kimi K3OPEN Moonshot AI	86.0	arXiv:2308.11462	2026-07
7	Grok 4.5 xAI	86.0	arXiv:2308.11462	2026-07
8	Muse Spark 1.2 Meta	85.3	arXiv:2308.11462	2026-08
9	GPT-5.6 Terra OpenAI	85.1	arXiv:2308.11462	2026-07
10	Muse Spark 1.1 Meta	85.0	arXiv:2308.11462	2026-07
11	Gemini 3.5 Flash Lite Google	84.1	arXiv:2308.11462	2026-07
12	GLM 5.2OPEN Zhipu AI	84.1	arXiv:2308.11462	2026-06
13	GPT-5.6 Luna OpenAI	84.0	arXiv:2308.11462	2026-07
14	Claude Sonnet 5 Anthropic	83.9	arXiv:2308.11462	2026-06
15	Qwen 3.8 MaxOPEN Alibaba	83.6	arXiv:2308.11462	2026-08
16	Inkling SmallOPEN Thinking Machines	83.0	arXiv:2308.11462	2026-07
17	InklingOPEN Thinking Machines	83.0	arXiv:2308.11462	2026-07
18	DeepSeek V4 Pro (0813)OPEN DeepSeek	82.4	arXiv:2308.11462	2026-08
19	Nemotron 3 UltraOPEN NVIDIA	82.1	arXiv:2308.11462	2026-06
20	Ling 3.0 FlashOPEN Ant Group	79.7	arXiv:2308.11462	2026-07
21	DeepSeek V4 FlashOPEN DeepSeek	77.7	arXiv:2308.11462	2026-07