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.

2026-10-20UPDATEPlatform updated: 52 models tracked across 42 independent benchmarks including CorpFin • 2026-10-20BENCHCorpFin added to benchmark suite — long credit agreement comprehension evaluations live • 2026-10-19MODELClaude Sonnet 5.5 and Opus 5.5 achieve 75.9% and 75.8% on Excel Modeling Benchmark v1.1 • 2026-10-19MODELTax Agent Bench updated: Claude Fable 5.1 (49.2%) and Gemini 4 Argon (44.6%) lead corporate tax research • 2026-10-18MODELGemini 4 Argon claims #1 rank on Finance Agent (v2) with 65.4% • 2026-10-13MODELGoogle launches Gemini 4 Argon with 1M-token output limit and native multimodal reasoning • 2026-10-06MODELAnthropic completes Claude 5.5 generation with Claude Haiku 5.5, Opus 5.5, and Sonnet 5.5 • 2026-10-20UPDATEPlatform updated: 52 models tracked across 42 independent benchmarks including CorpFin • 2026-10-20BENCHCorpFin added to benchmark suite — long credit agreement comprehension evaluations live • 2026-10-19MODELClaude Sonnet 5.5 and Opus 5.5 achieve 75.9% and 75.8% on Excel Modeling Benchmark v1.1 • 2026-10-19MODELTax Agent Bench updated: Claude Fable 5.1 (49.2%) and Gemini 4 Argon (44.6%) lead corporate tax research • 2026-10-18MODELGemini 4 Argon claims #1 rank on Finance Agent (v2) with 65.4% • 2026-10-13MODELGoogle launches Gemini 4 Argon with 1M-token output limit and native multimodal reasoning • 2026-10-06MODELAnthropic completes Claude 5.5 generation with Claude Haiku 5.5, Opus 5.5, and Sonnet 5.5 •

Tier 2agentTasks

SWE-bench

Software engineering benchmark evaluating resolution of real-world GitHub issues.

2,294 tasks

22 models evaluated

Updated 2026

arXiv:2310.06770

View Source →Submit your score Methodology

97.0%

Top Score

Top model

Claude Opus 5

Model Rankings

financebenchmark.ai · SWE-bench

93.4%

🥇

AnthropicClaude Opus 597.0%

🥈

AnthropicClaude Sonnet 579.6%

🥉

OpenAIGPT-5.6 Sol96.2%

4

DeepSeekDeepSeek V4 FlashOPEN88.8%

5

xAIGrok 4.695.6%

6

Moonshot AIKimi K3OPEN93.4%

7

AnthropicClaude Fable 595.0%

8

DeepSeekDeepSeek V4 Pro (0813)OPEN96.4%

9

GoogleGemini 3.5 Flash Lite75.0%

10

GoogleGemini 3.6 Flash74.8%

11

Zhipu AIGLM 5.2OPEN82.8%

12

OpenAIGPT-5.6 Luna93.0%

13

OpenAIGPT-5.6 Terra75.2%

14

xAIGrok 4.586.6%

15

Thinking MachinesInklingOPEN77.6%

16

Thinking MachinesInkling SmallOPEN82.2%

17

Moonshot AIKimi K2.7 CodeOPEN78.2%

18

Ant GroupLing 3.0 FlashOPEN65.2%

19

MetaMuse Spark 1.182.0%

20

MetaMuse Spark 1.286.6%

21

NVIDIANemotron 3 UltraOPEN69.0%

22

AlibabaQwen 3.8 MaxOPEN85.3%

Full Results Table

#	Model	Overall	Source	Date
🥇	Claude Opus 5 Anthropic	97.0	arXiv:2310.06770	2026-07
🥈	DeepSeek V4 Pro (0813)OPEN DeepSeek	96.4	arXiv:2310.06770	2026-08
🥉	GPT-5.6 Sol OpenAI	96.2	arXiv:2310.06770	2026-07
4	Grok 4.6 xAI	95.6	arXiv:2310.06770	2026-08
5	Claude Fable 5 Anthropic	95.0	arXiv:2310.06770	2026-06
6	Kimi K3OPEN Moonshot AI	93.4	arXiv:2310.06770	2026-07
7	GPT-5.6 Luna OpenAI	93.0	arXiv:2310.06770	2026-07
8	DeepSeek V4 FlashOPEN DeepSeek	88.8	arXiv:2310.06770	2026-07
9	Grok 4.5 xAI	86.6	arXiv:2310.06770	2026-07
10	Muse Spark 1.2 Meta	86.6	arXiv:2310.06770	2026-08
11	Qwen 3.8 MaxOPEN Alibaba	85.3	arXiv:2310.06770	2026-08
12	GLM 5.2OPEN Zhipu AI	82.8	arXiv:2310.06770	2026-06
13	Inkling SmallOPEN Thinking Machines	82.2	arXiv:2310.06770	2026-07
14	Muse Spark 1.1 Meta	82.0	arXiv:2310.06770	2026-07
15	Claude Sonnet 5 Anthropic	79.6	arXiv:2310.06770	2026-06
16	Kimi K2.7 CodeOPEN Moonshot AI	78.2	arXiv:2310.06770	2026-06
17	InklingOPEN Thinking Machines	77.6	arXiv:2310.06770	2026-07
18	GPT-5.6 Terra OpenAI	75.2	arXiv:2310.06770	2026-07
19	Gemini 3.5 Flash Lite Google	75.0	arXiv:2310.06770	2026-07
20	Gemini 3.6 Flash Google	74.8	arXiv:2310.06770	2026-07
21	Nemotron 3 UltraOPEN NVIDIA	69.0	arXiv:2310.06770	2026-06
22	Ling 3.0 FlashOPEN Ant Group	65.2	arXiv:2310.06770	2026-07

Contents.md

All evaluated models

Claude Opus 5

97.0%

Claude Sonnet 5

79.6%

GPT-5.6 Sol

96.2%

DeepSeek V4 Flash

88.8%

Grok 4.6

95.6%

Kimi K3

93.4%

Claude Fable 5

95.0%

DeepSeek V4 Pro (0813)

96.4%

Gemini 3.5 Flash Lite

75.0%

Gemini 3.6 Flash

74.8%

GLM 5.2

82.8%

GPT-5.6 Luna

93.0%

GPT-5.6 Terra

75.2%

Grok 4.5

86.6%

Inkling

77.6%

Inkling Small

82.2%

Kimi K2.7 Code

78.2%

Ling 3.0 Flash

65.2%

Muse Spark 1.1

82.0%

Muse Spark 1.2

86.6%

Nemotron 3 Ultra

69.0%

Qwen 3.8 Max

85.3%