Agentic · tools

AA Tau3 Banking leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

As of 2026-10-07, the highest measured score on AA Tau3 Banking is 50.7% by Grok 4.6. 135 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Grok 4.6	50.7%	measured
2	Muse Spark 1.3	50.5%	measured
3	GLM-5.3	50.3%	measured
4	Claude Opus 5.5	49.3%	estimated ± 3.0 pp, low confidence
5	Grok 4.7	48.2%	estimated ± 3.0 pp, low confidence
6	GLM-5.2	48.0%	estimated ± 3.0 pp, low confidence
7	Qwen3.8-27B	48.0%	measured
8	Gemini 3.7 Flash	47.7%	estimated ± 4.4 pp, low confidence
9	DeepSeek V4 Pro 0813	47.5%	estimated ± 5.8 pp, medium confidence
10	Gemini 4 Argon	47.5%	estimated ± 5.8 pp, medium confidence
11	Ling 3.1 Flash	47.5%	estimated ± 5.8 pp, medium confidence
12	MiMo-V2.6-Flash	47.5%	estimated ± 5.8 pp, medium confidence
13	MiMo-V2.6-Pro	47.5%	estimated ± 5.8 pp, medium confidence
14	Qwen3.8-Flash-Next	47.5%	estimated ± 5.8 pp, medium confidence
15	Qwen3.8 Max Preview	47.5%	estimated ± 5.8 pp, medium confidence
16	Muse Spark 1.2	47.5%	estimated ± 5.8 pp, medium confidence
17	Hy4 preview	47.4%	estimated ± 10.7 pp, low confidence
18	Claude Sonnet 5	47.4%	estimated ± 5.8 pp, medium confidence
19	GPT-5.6 Luna	47.4%	estimated ± 5.8 pp, medium confidence
20	Claude Opus 4.8	47.4%	estimated ± 5.8 pp, medium confidence
21	DeepSeek V4 Flash 0731	47.2%	estimated ± 5.8 pp, medium confidence
22	GPT-6 Luna	47.2%	estimated ± 5.8 pp, medium confidence
23	Claude Fable 5.1	47.2%	measured
24	GLM-5.3-Flash	47.2%	measured
25	GPT-5.5	47.1%	estimated ± 3.0 pp, low confidence
26	Claude Sonnet 5.5	47.0%	estimated ± 4.4 pp, medium confidence
27	Mistral Large 4	46.9%	estimated ± 5.8 pp, medium confidence
28	Claude Opus 4.7 (Adaptive)	46.9%	estimated ± 3.0 pp, low confidence
29	DeepSeek V4.1 Flash	46.8%	estimated ± 3.0 pp, low confidence
30	GPT-6 Sol	46.1%	estimated ± 3.0 pp, low confidence
31	Kimi K3	46.0%	measured
32	Claude Opus 5	45.9%	estimated ± 4.4 pp, medium confidence
33	GPT-5.6 Terra	45.6%	estimated ± 3.0 pp, low confidence
34	Gemini 3.8 Flash	44.9%	measured
35	GPT-6.1 Sol	44.6%	estimated ± 4.4 pp, medium confidence
36	Step 5 Preview	44.3%	estimated ± 3.0 pp, low confidence
37	Claude Fable 5	44.2%	estimated ± 4.4 pp, medium confidence
38	GPT-5.6 Sol	44.1%	estimated ± 3.0 pp, low confidence
39	Grok 4.5	42.5%	estimated ± 5.8 pp, medium confidence
40	Atria Dawn Preview	41.4%	estimated ± 10.2 pp, low confidence
41	GPT-6 Astra	41.4%	measured
42	GPT-5.5 Pro	38.4%	estimated ± 10.2 pp, low confidence
43	GPT-5.4 Pro	37.4%	estimated ± 10.2 pp, low confidence
44	Claude Mythos 5	35.9%	estimated ± 10.2 pp, low confidence
45	Qwen3.8 Max	35.3%	estimated ± 14.6 pp, low confidence
46	Claude Haiku 5.5	34.8%	estimated ± 7.1 pp, medium confidence
47	Ornith-1.5-397B	34.3%	estimated ± 10.2 pp, low confidence
48	Claude Opus 4.6	31.3%	estimated ± 10.2 pp, low confidence
49	Claude Opus 4.7	31.0%	estimated ± 14.6 pp, low confidence
50	dots3-note Preview	30.9%	estimated ± 10.2 pp, low confidence
51	Inkling	29.1%	measured
52	Ling 3.0 Flash	28.0%	measured
53	Gemini 3.5 Flash	26.0%	estimated ± 5.8 pp, medium confidence
54	Beam	25.7%	estimated ± 10.2 pp, low confidence
55	Agents-A1	24.3%	estimated ± 10.2 pp, low confidence
56	Claude Sonnet 4.6	24.1%	estimated ± 14.6 pp, low confidence
57	Muse Glimmer 30B	23.5%	measured
58	Gemini 3.6 Flash	22.1%	estimated ± 5.8 pp, medium confidence
59	GPT-5.4	22.0%	estimated ± 5.8 pp, medium confidence
60	Hy3 Preview	22.0%	estimated ± 5.8 pp, medium confidence
61	Muse Spark 1.1	22.0%	estimated ± 5.8 pp, medium confidence
62	Apodex 1.1	22.0%	estimated ± 5.8 pp, medium confidence
63	Apodex 1.1 Mini	22.0%	estimated ± 5.8 pp, medium confidence
64	Quasar 438B	22.0%	estimated ± 5.8 pp, medium confidence
65	A.X K2	22.0%	estimated ± 5.8 pp, medium confidence
66	Celeris-1	22.0%	estimated ± 5.8 pp, low confidence
67	Command A+	22.0%	estimated ± 5.8 pp, low confidence
68	DeepSeek V3	22.0%	estimated ± 5.8 pp, low confidence
69	DeepSeek V3 0324	22.0%	estimated ± 5.8 pp, low confidence
70	Gemini 2.5 Pro	22.0%	estimated ± 5.8 pp, low confidence
71	Gemini 3.1 Pro	22.0%	estimated ± 5.8 pp, medium confidence
72	Gemini 3.5 Flash-Lite	22.0%	estimated ± 5.8 pp, medium confidence
73	Gemma 3 27B	22.0%	estimated ± 5.8 pp, low confidence
74	Gemma 4 12B	22.0%	estimated ± 5.8 pp, low confidence
75	Gemma 4 26B A4B	22.0%	estimated ± 5.8 pp, low confidence
76	Gemma 4 31B	22.0%	estimated ± 5.8 pp, low confidence
77	Gemma 4 E2B	22.0%	estimated ± 5.8 pp, low confidence
78	Gemma 4 E4B	22.0%	estimated ± 5.8 pp, low confidence
79	GLM-4.7	22.0%	estimated ± 5.8 pp, medium confidence
80	GLM-5.1	22.0%	estimated ± 5.8 pp, medium confidence
81	GPT-4.1 mini	22.0%	estimated ± 5.8 pp, low confidence
82	GPT-4.1 nano	22.0%	estimated ± 5.8 pp, low confidence
83	GPT-4o	22.0%	estimated ± 5.8 pp, low confidence
84	GPT-4o mini	22.0%	estimated ± 5.8 pp, low confidence
85	GPT-5.1	22.0%	estimated ± 5.8 pp, medium confidence
86	GPT-5.4 mini	22.0%	estimated ± 5.8 pp, medium confidence
87	GPT-5.4 nano	22.0%	estimated ± 5.8 pp, medium confidence
88	GPT-5 (high)	22.0%	estimated ± 5.8 pp, medium confidence
89	GPT-OSS 120B	22.0%	estimated ± 5.8 pp, low confidence
90	GPT-OSS 20B	22.0%	estimated ± 5.8 pp, low confidence
91	Granite 4.2 30B	22.0%	estimated ± 5.8 pp, low confidence
92	Granite 4.2 3B	22.0%	estimated ± 5.8 pp, low confidence
93	Granite 4.2 8B	22.0%	estimated ± 5.8 pp, low confidence
94	Grok 4.3	22.0%	estimated ± 5.8 pp, medium confidence
95	Hy3	22.0%	estimated ± 5.8 pp, medium confidence
96	Inkling-Small	22.0%	estimated ± 5.8 pp, medium confidence
97	K-Exaone	22.0%	estimated ± 5.8 pp, low confidence
98	Kimi K2.6	22.0%	estimated ± 5.8 pp, medium confidence
99	Kimi K2.5	22.0%	estimated ± 5.8 pp, medium confidence
100	Kimi K2.5 (Reasoning)	22.0%	estimated ± 5.8 pp, medium confidence
101	Kimi K2.7 Code	22.0%	estimated ± 5.8 pp, medium confidence
102	LFM2.5-2.6B	22.0%	estimated ± 5.8 pp, low confidence
103	Ling 2.6 Flash	22.0%	estimated ± 5.8 pp, low confidence
104	Ling 3.0 Flash FP8	22.0%	estimated ± 5.8 pp, medium confidence
105	Ling 3.0 Flash VL	22.0%	estimated ± 5.8 pp, medium confidence
106	Ling 3.0 Tiny	22.0%	estimated ± 5.8 pp, low confidence
107	Llama 4 Maverick	22.0%	estimated ± 5.8 pp, low confidence
108	Llama 4 Scout	22.0%	estimated ± 5.8 pp, low confidence
109	Mercury 2.5	22.0%	estimated ± 5.8 pp, low confidence
110	MiMo-V2.5-Pro	22.0%	estimated ± 5.8 pp, medium confidence
111	MiMo-V2-Flash	22.0%	estimated ± 5.8 pp, low confidence
112	MiniCPM5-2B	22.0%	estimated ± 5.8 pp, low confidence
113	MiniMax M2.7	22.0%	estimated ± 5.8 pp, medium confidence
114	Mistral Large 3	22.0%	estimated ± 5.8 pp, low confidence
115	Mistral Medium 3.5 128B	22.0%	estimated ± 5.8 pp, low confidence
116	Mistral Small 4	22.0%	estimated ± 5.8 pp, low confidence
117	Mistral Small 4 (Reasoning)	22.0%	estimated ± 5.8 pp, low confidence
118	Muse Spark	22.0%	estimated ± 5.8 pp, medium confidence
119	Nemotron 3.5 Lightning 30B A3B NVFP4	22.0%	estimated ± 5.8 pp, low confidence
120	Nemotron 3 Nano 30B	22.0%	estimated ± 5.8 pp, low confidence
121	Nemotron 3 Nano Omni 30B A3B	22.0%	estimated ± 5.8 pp, low confidence
122	Nemotron 3 Super 100B	22.0%	estimated ± 5.8 pp, low confidence
123	North Mini Code	22.0%	estimated ± 5.8 pp, low confidence
124	Qwen3.5-122B-A10B	22.0%	estimated ± 5.8 pp, medium confidence
125	Qwen3.6-27B	22.0%	estimated ± 5.8 pp, medium confidence
126	Qwen3.6-35B-A3B	22.0%	estimated ± 5.8 pp, medium confidence
127	Qwen3.6 Plus	22.0%	estimated ± 5.8 pp, medium confidence
128	Qwen3.7 Max	22.0%	estimated ± 5.8 pp, medium confidence
129	Qwen3.7 Plus	22.0%	estimated ± 5.8 pp, low confidence
130	Solar Pro 3	22.0%	estimated ± 5.8 pp, low confidence
131	Solar Pro 4	22.0%	estimated ± 5.8 pp, medium confidence
132	Step 3.7 Flash	22.0%	estimated ± 5.8 pp, medium confidence
133	Trinity-Large-Preview	22.0%	estimated ± 5.8 pp, low confidence
134	Trinity-Large-Thinking	22.0%	estimated ± 5.8 pp, low confidence
135	Ultravox v0.6 Llama 3.3 70B	22.0%	estimated ± 5.8 pp, low confidence
136	Ornith-1.5-35B-A3B	19.9%	estimated ± 10.2 pp, low confidence
137	Agents-A1-4B	19.6%	estimated ± 10.2 pp, low confidence
138	GPT-5.2	19.2%	estimated ± 10.2 pp, low confidence
139	Qwen3.5 397B	17.9%	estimated ± 10.2 pp, low confidence
140	Qwen3.5-27B	17.6%	estimated ± 10.2 pp, low confidence
141	Qwen3.5-35B-A3B	17.6%	estimated ± 10.2 pp, low confidence
142	Ornith-1.5-9B	16.5%	estimated ± 10.2 pp, low confidence
143	LongCat-Flash-Lite-Sparse	15.5%	estimated ± 10.2 pp, low confidence
144	MiniMax M3	15.3%	measured
145	Nemotron 3 Ultra	14.2%	measured
146	Solar Open 2	13.0%	estimated ± 11.8 pp, low confidence
147	LLaDA2.2-flash	8.2%	estimated ± 11.8 pp, low confidence
148	Claude Opus 4.5	7.0%	estimated ± 11.8 pp, low confidence
149	GLM-5	4.4%	estimated ± 11.8 pp, low confidence

AA Tau3 Banking leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General