Agents · Benchmark profile

τ³-Banking

Fintech customer-support agent benchmark requiring knowledge-base navigation and multi-step tool use on banking workflows.

Ranking weighted3Active57 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on τ³-Banking

Leaderboard · 57 models

Kimi K3 leads at 33.4%, followed by GPT-5.6 Sol (33.0%) and Grok 4.5 (32.6%).

Ranking · 1.2%Agents3

#1Closed

Kimi K3

Moonshot AI

kimi-k3

33.4%

Overall #3 · Context 1M tokens

#2Closed

GPT-5.6 Sol

OpenAI

gpt-5.6-sol

33.0%

Overall #4 · Context 1.05M tokens

#3Closed

Grok 4.5

xAI

grok-4.5

32.6%

Overall #6 · Context 500K tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

Moonshot AIOpenAIxAIAnthropicZ.AIDeepSeekGoogleMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Kimi K3	#1	Moonshot AI	33.4%
GPT-5.6 Sol	#2	OpenAI	33.0%
Grok 4.5	#3	xAI	32.6%
GPT-5.6 Terra	#4	OpenAI	31.8%
GPT-5.5	#5	OpenAI	31.3%
Claude Sonnet 4.6	#6	Anthropic	30.5%
GPT-5.4	#7	OpenAI	30.3%
Claude Opus 4.7	#8	Anthropic	28.9%
Claude Sonnet 5	#9	Anthropic	28.2%
Claude Opus 4.8	#10	Anthropic	27.6%
GPT-5.6 Luna	#11	OpenAI	27.2%
Claude Fable 5	#12	Anthropic	26.8%
GLM-5.2	#13	Z.AI	26.8%
DeepSeek V4 Pro	#14	DeepSeek	25.8%
Gemini 3.5 Flash	#15	Google	25.4%
Muse Spark 1.1	#16	Meta	25.2%
DeepSeek V4 Flash	#17	DeepSeek	22.9%
GPT-5.4 mini	#18	OpenAI	21.4%
GPT-5.4 nano	#19	OpenAI	21.0%
Kimi K2.6	#20	Moonshot AI	20.6%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Kimi K3	#1	Moonshot AI	33.4%
GPT-5.6 Sol	#2	OpenAI	33.0%
Grok 4.5	#3	xAI	32.6%
GPT-5.6 Terra	#4	OpenAI	31.8%
GPT-5.5	#5	OpenAI	31.3%
Claude Sonnet 4.6	#6	Anthropic	30.5%
GPT-5.4	#7	OpenAI	30.3%
Claude Opus 4.7	#8	Anthropic	28.9%
Claude Sonnet 5	#9	Anthropic	28.2%
Claude Opus 4.8	#10	Anthropic	27.6%
GPT-5.6 Luna	#11	OpenAI	27.2%
Claude Fable 5	#12	Anthropic	26.8%
GLM-5.2	#13	Z.AI	26.8%
DeepSeek V4 Pro	#14	DeepSeek	25.8%
Gemini 3.5 Flash	#15	Google	25.4%
Muse Spark 1.1	#16	Meta	25.2%
DeepSeek V4 Flash	#17	DeepSeek	22.9%
GPT-5.4 mini	#18	OpenAI	21.4%
GPT-5.4 nano	#19	OpenAI	21.0%
Kimi K2.6	#20	Moonshot AI	20.6%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
xAI	1	32.6%
Anthropic	5	28.4%
OpenAI	7	28.0%
Moonshot AI	2	27.0%
Z.AI	1	26.8%
Google	1	25.4%
Meta	1	25.2%
DeepSeek	2	24.3%

Benchmark score table (57 models)

One best score per model · higher is better

τ³-Banking leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Kimi K3 kimi-k3	Moonshot AI	closed	33.4%
#2	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	33.0%
#3	Grok 4.5 grok-4.5	xAI	closed	32.6%
#4	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	31.8%
#5	GPT-5.5 gpt-5.5	OpenAI	closed	31.3%
#6	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	30.5%
#7	GPT-5.4 gpt-5.4	OpenAI	closed	30.3%
#8	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	28.9%
#9	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	28.2%
#10	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	27.6%
#11	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	27.2%
#12	Claude Fable 5 claude-fable-5	Anthropic	closed	26.8%
#13	GLM-5.2 glm-5-2	Z.AI	open	26.8%
#14	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	25.8%
#15	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	25.4%
#16	Muse Spark 1.1 muse-spark-1-1	Meta	closed	25.2%
#17	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	22.9%
#18	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	21.4%
#19	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	21.0%
#20	Kimi K2.6 kimi-k2.6	Moonshot AI	open	20.6%
#21	GPT-5 gpt-5	OpenAI	closed	19.6%
#22	Muse Spark muse-spark	Meta	closed	19.6%
#23	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	19.0%
#24	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	18.8%
#25	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	18.1%
#26	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	17.9%
#27	Gemini 3 Flash gemini-3-flash	Google	closed	17.5%
#28	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	16.5%
#29	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	16.5%
#30	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	15.9%
#31	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	15.3%
#32	Gemma 4 31B gemma-4-31b	Google	open	15.1%
#33	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	14.4%
#34	Kimi K2.5 kimi-k2.5	Moonshot AI	—	14.2%
#35	GPT-5.1 gpt.5.1	OpenAI	closed	14.0%
#36	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	13.8%
#37	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	13.8%
#38	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	13.6%
#39	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	13.4%
#40	MiniMax M3 MiniMax-M3	MiniMax	—	13.0%
#41	Grok 4.3 grok-4.3	xAI	closed	12.2%
#42	Gemma 4 26B gemma-4-26b-a4b	Google	open	11.8%
#43	GLM-5.1 glm-5-1	Z.AI	open	11.5%
#44	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	10.9%
#45	GLM-4.6 glm-4.6	Z.AI	open	10.5%
#46	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	10.1%
#47	GLM-4.7 glm-4.7	Z.AI	open	9.7%
#48	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	9.3%
#49	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	9.1%
#50	MiniMax M2.7 minimax-m2-7	MiniMax	open	8.9%
#51	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	8.7%
#52	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	8.7%
#53	Nova 2 Pro nova-2-pro-preview	Amazon	closed	7.6%
#54	MiMo-V2.5 mimo-v2.5	Xiaomi	open	6.6%
#55	Command A+ command-a-plus	Cohere	open	5.8%
#56	Mistral Large 3 mistral-large-2512	Mistral AI	open	5.8%
#57	Mistral Small 4 mistral-small-2603	Mistral AI	open	5.2%

The top of this snapshot is led by Kimi K3 at 33.4%; third place is 0.82 points behind. The top-10 spread is 5.8 points.

Compare top models

Kimi K3 vs GPT-5.6 Sol GPT-5.6 Sol vs Grok 4.5 Grok 4.5 vs GPT-5.6 Terra GPT-5.6 Terra vs GPT-5.5

About τ³-Banking

Definition and scoring

Organisation: Sierra / Artificial Analysis
Category: Agents
Version: 3
Direction: higher is better
Ranking use: Weighted 1.2%
Contamination risk: Low

Fintech customer-support agent benchmark requiring knowledge-base navigation and multi-step tool use on banking workflows. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.

Open benchmark source ↗

FAQ

Quick answers

What does τ³-Banking measure?

Fintech customer-support agent benchmark requiring knowledge-base navigation and multi-step tool use on banking workflows.

Which model scores highest on τ³-Banking?

Kimi K3 by Moonshot AI currently leads with 33.4%.

How many models are evaluated on τ³-Banking?

57 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 1.2% of its capability category. Kimi K3 is overall #3.

Terminal-Bench

2.1 · 80 results · Weighted 1.6%

Berkeley Function-Calling Leaderboard

V3 · 2 results · Not Weighted

τ-bench

Rolling · 37 results · Weighted 1.6%

GAIA

Rolling · 13 results · Weighted 1.6%

BrowseComp

Rolling · 24 results · Weighted 1.6%

OSWorld

Rolling · 11 results · Weighted 1.6%

Agents · Benchmark profile

τ³-Banking

Fintech customer-support agent benchmark requiring knowledge-base navigation and multi-step tool use on banking workflows.

Ranking weighted3Active57 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on τ³-Banking

Leaderboard · 57 models

Kimi K3 leads at 33.4%, followed by GPT-5.6 Sol (33.0%) and Grok 4.5 (32.6%).

Ranking · 1.2%Agents3

#1Closed

Kimi K3

Moonshot AI

kimi-k3

33.4%

Overall #3 · Context 1M tokens

#2Closed

GPT-5.6 Sol

OpenAI

gpt-5.6-sol

33.0%

Overall #4 · Context 1.05M tokens

#3Closed

Grok 4.5

xAI

grok-4.5

32.6%

Overall #6 · Context 500K tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

Moonshot AIOpenAIxAIAnthropicZ.AIDeepSeekGoogleMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Kimi K3	#1	Moonshot AI	33.4%
GPT-5.6 Sol	#2	OpenAI	33.0%
Grok 4.5	#3	xAI	32.6%
GPT-5.6 Terra	#4	OpenAI	31.8%
GPT-5.5	#5	OpenAI	31.3%
Claude Sonnet 4.6	#6	Anthropic	30.5%
GPT-5.4	#7	OpenAI	30.3%
Claude Opus 4.7	#8	Anthropic	28.9%
Claude Sonnet 5	#9	Anthropic	28.2%
Claude Opus 4.8	#10	Anthropic	27.6%
GPT-5.6 Luna	#11	OpenAI	27.2%
Claude Fable 5	#12	Anthropic	26.8%
GLM-5.2	#13	Z.AI	26.8%
DeepSeek V4 Pro	#14	DeepSeek	25.8%
Gemini 3.5 Flash	#15	Google	25.4%
Muse Spark 1.1	#16	Meta	25.2%
DeepSeek V4 Flash	#17	DeepSeek	22.9%
GPT-5.4 mini	#18	OpenAI	21.4%
GPT-5.4 nano	#19	OpenAI	21.0%
Kimi K2.6	#20	Moonshot AI	20.6%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Kimi K3	#1	Moonshot AI	33.4%
GPT-5.6 Sol	#2	OpenAI	33.0%
Grok 4.5	#3	xAI	32.6%
GPT-5.6 Terra	#4	OpenAI	31.8%
GPT-5.5	#5	OpenAI	31.3%
Claude Sonnet 4.6	#6	Anthropic	30.5%
GPT-5.4	#7	OpenAI	30.3%
Claude Opus 4.7	#8	Anthropic	28.9%
Claude Sonnet 5	#9	Anthropic	28.2%
Claude Opus 4.8	#10	Anthropic	27.6%
GPT-5.6 Luna	#11	OpenAI	27.2%
Claude Fable 5	#12	Anthropic	26.8%
GLM-5.2	#13	Z.AI	26.8%
DeepSeek V4 Pro	#14	DeepSeek	25.8%
Gemini 3.5 Flash	#15	Google	25.4%
Muse Spark 1.1	#16	Meta	25.2%
DeepSeek V4 Flash	#17	DeepSeek	22.9%
GPT-5.4 mini	#18	OpenAI	21.4%
GPT-5.4 nano	#19	OpenAI	21.0%
Kimi K2.6	#20	Moonshot AI	20.6%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
xAI	1	32.6%
Anthropic	5	28.4%
OpenAI	7	28.0%
Moonshot AI	2	27.0%
Z.AI	1	26.8%
Google	1	25.4%
Meta	1	25.2%
DeepSeek	2	24.3%

Benchmark score table (57 models)

One best score per model · higher is better

τ³-Banking leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Kimi K3 kimi-k3	Moonshot AI	closed	33.4%
#2	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	33.0%
#3	Grok 4.5 grok-4.5	xAI	closed	32.6%
#4	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	31.8%
#5	GPT-5.5 gpt-5.5	OpenAI	closed	31.3%
#6	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	30.5%
#7	GPT-5.4 gpt-5.4	OpenAI	closed	30.3%
#8	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	28.9%
#9	Claude Sonnet 5 claude-sonnet-5	Anthropic	closed	28.2%
#10	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	27.6%
#11	GPT-5.6 Luna gpt-5.6-luna	OpenAI	closed	27.2%
#12	Claude Fable 5 claude-fable-5	Anthropic	closed	26.8%
#13	GLM-5.2 glm-5-2	Z.AI	open	26.8%
#14	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	25.8%
#15	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	25.4%
#16	Muse Spark 1.1 muse-spark-1-1	Meta	closed	25.2%
#17	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	22.9%
#18	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	21.4%
#19	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	21.0%
#20	Kimi K2.6 kimi-k2.6	Moonshot AI	open	20.6%
#21	GPT-5 gpt-5	OpenAI	closed	19.6%
#22	Muse Spark muse-spark	Meta	closed	19.6%
#23	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	19.0%
#24	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	18.8%
#25	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	18.1%
#26	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	17.9%
#27	Gemini 3 Flash gemini-3-flash	Google	closed	17.5%
#28	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	16.5%
#29	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	16.5%
#30	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	15.9%
#31	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	15.3%
#32	Gemma 4 31B gemma-4-31b	Google	open	15.1%
#33	Mistral Medium 3.5 mistral-medium-3-5	Mistral AI	open	14.4%
#34	Kimi K2.5 kimi-k2.5	Moonshot AI	—	14.2%
#35	GPT-5.1 gpt.5.1	OpenAI	closed	14.0%
#36	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	13.8%
#37	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	13.8%
#38	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	13.6%
#39	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	13.4%
#40	MiniMax M3 MiniMax-M3	MiniMax	—	13.0%
#41	Grok 4.3 grok-4.3	xAI	closed	12.2%
#42	Gemma 4 26B gemma-4-26b-a4b	Google	open	11.8%
#43	GLM-5.1 glm-5-1	Z.AI	open	11.5%
#44	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	10.9%
#45	GLM-4.6 glm-4.6	Z.AI	open	10.5%
#46	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	10.1%
#47	GLM-4.7 glm-4.7	Z.AI	open	9.7%
#48	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	9.3%
#49	Claude Haiku 4.5 claude-haiku-4-5-20251001	Anthropic	closed	9.1%
#50	MiniMax M2.7 minimax-m2-7	MiniMax	open	8.9%
#51	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	8.7%
#52	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	8.7%
#53	Nova 2 Pro nova-2-pro-preview	Amazon	closed	7.6%
#54	MiMo-V2.5 mimo-v2.5	Xiaomi	open	6.6%
#55	Command A+ command-a-plus	Cohere	open	5.8%
#56	Mistral Large 3 mistral-large-2512	Mistral AI	open	5.8%
#57	Mistral Small 4 mistral-small-2603	Mistral AI	open	5.2%

The top of this snapshot is led by Kimi K3 at 33.4%; third place is 0.82 points behind. The top-10 spread is 5.8 points.

Compare top models

Kimi K3 vs GPT-5.6 Sol GPT-5.6 Sol vs Grok 4.5 Grok 4.5 vs GPT-5.6 Terra GPT-5.6 Terra vs GPT-5.5

About τ³-Banking

Definition and scoring

Organisation: Sierra / Artificial Analysis
Category: Agents
Version: 3
Direction: higher is better
Ranking use: Weighted 1.2%
Contamination risk: Low

Open benchmark source ↗

FAQ

Quick answers

What does τ³-Banking measure?

Fintech customer-support agent benchmark requiring knowledge-base navigation and multi-step tool use on banking workflows.

Which model scores highest on τ³-Banking?

Kimi K3 by Moonshot AI currently leads with 33.4%.

How many models are evaluated on τ³-Banking?

57 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

Yes. This family is ranking-weighted at 1.2% of its capability category. Kimi K3 is overall #3.

Terminal-Bench

2.1 · 80 results · Weighted 1.6%

Berkeley Function-Calling Leaderboard

V3 · 2 results · Not Weighted

τ-bench

Rolling · 37 results · Weighted 1.6%

GAIA

Rolling · 13 results · Weighted 1.6%

BrowseComp

Rolling · 24 results · Weighted 1.6%

OSWorld

Rolling · 11 results · Weighted 1.6%

Leaderboard · 57 models

Kimi K3

GPT-5.6 Sol

Grok 4.5

See the field, not just the table.

Benchmark score table (57 models)

Compare top models

Definition and scoring

Quick answers

What does τ³-Banking measure?

Which model scores highest on τ³-Banking?

How many models are evaluated on τ³-Banking?

Does this affect overall Lumina rank?

More Agents benchmarks

Leaderboard · 57 models

Kimi K3

GPT-5.6 Sol

Grok 4.5

See the field, not just the table.

Benchmark score table (57 models)

Compare top models

Definition and scoring

Quick answers

What does τ³-Banking measure?

Which model scores highest on τ³-Banking?

How many models are evaluated on τ³-Banking?

Does this affect overall Lumina rank?

More Agents benchmarks