Workflow automation

Automation platforms that chain apps and AI steps into workflows that run on triggers, and agent builders that let AI decide and act across your tools.

Models are ranked by AutomationBench-AA (Artificial Analysis). Tools inherit the tier of the best model they offer. How we rank.

Best overallSGemini 4 Argon77.5% on AutomationBench-AA (Artificial Analysis)
Best valueSGemini 4 ArgonTier S for $4/M tokens
Best free startAChatGPTFree tier · from $20/mo · via GPT-6 Astra
Best local (≤24 GB GPU)CQwen3.8-27B~17 GB VRAM · 48.2% on AutomationBench-AA (Artificial Analysis)

Tier: how close to the best result on the task's main public benchmark
Outlined: a tool inherits the tier of the best model it offers
No public result yet
Cloud · Local: runs on your own hardware

Models

The engines themselves. Use them through an API or inside the apps on the right.

Cloud

Gemini 4 ArgonGoogle · $4/M tokens77.5% on AutomationBench-AA (Artificial Analysis)$A

Claude Sonnet 5.5Anthropic · $4/M tokens71.8% on AutomationBench-AA (Artificial Analysis)A Muse WonderMeta · $8/M tokens71.3% on AutomationBench-AA (Artificial Analysis)A

Claude Opus 5.5Anthropic · $8/M tokens69.5% on AutomationBench-AA (Artificial Analysis)A

GPT-6 AstraOpenAI · $20/M tokens68.5% on AutomationBench-AA (Artificial Analysis)B

Grok 4.7xAI · $3/M tokens65.6% on AutomationBench-AA (Artificial Analysis)$B

GPT-6.1 SolOpenAI · $4/M tokens64.9% on AutomationBench-AA (Artificial Analysis)B

Gemini 3.8 FlashGoogle · $1.5/M tokens59.9% on AutomationBench-AA (Artificial Analysis)$B

Mistral Large 4 (preview)Mistral AI · $2.06/M tokens59.9% on AutomationBench-AA (Artificial Analysis)C

Claude Fable 5.1Anthropic · $20/M tokens59.4% on AutomationBench-AA (Artificial Analysis)C Qwen3.8-MaxAlibaba (Qwen) · $3/M tokens56.2% on AutomationBench-AA (Artificial Analysis)C

GPT-6 LunaOpenAI · $0.2/M tokens53.2% on AutomationBench-AA (Artificial Analysis)$C

Claude Haiku 5.5Anthropic · $0.2/M tokens35.4% on AutomationBench-AA (Artificial Analysis)

Local

DeepSeek-V4.1-Flash-1017DeepSeek · ~320 GB VRAM66.6% on AutomationBench-AA (Artificial Analysis)$B

GLM-5.3Z.ai · ~425 GB VRAM62.2% on AutomationBench-AA (Artificial Analysis)B

GLM-5.3-FlashZ.ai · ~185 GB VRAM60.4% on AutomationBench-AA (Artificial Analysis)$C

MiMo-V2.6-ProXiaomi · ~585 GB VRAM58.6% on AutomationBench-AA (Artificial Analysis)C

Kimi K3Moonshot AI · ~1600 GB VRAM58.3% on AutomationBench-AA (Artificial Analysis)C

Step 5 PreviewStepFun · ~346 GB VRAM51% on AutomationBench-AA (Artificial Analysis)C

Muse Spark Open (816B-A42B)Meta · ~470 GB VRAM49.7% on AutomationBench-AA (Artificial Analysis)C

Beam 1.0 (501B-A23B)Reflection AI · ~287 GB VRAM48.5% on AutomationBench-AA (Artificial Analysis)C

Qwen3.8-27BAlibaba (Qwen) · ~17 GB VRAM48.2% on AutomationBench-AA (Artificial Analysis)$C

InklingThinking Machines Lab · ~557 GB VRAM5% on AutomationBench-AA (Artificial Analysis)C

K2-Horizon-375B-A23BMBZUAI Institute of Foundation Models · ~218 GB VRAM37.2% on AutomationBench-AA (Artificial Analysis)C

Nemotron 3.5 Super VL 120B-A12BNVIDIA · ~72 GB VRAM14.8% on AutomationBench-AA (Artificial Analysis)

Automation platforms

Connect apps and AI steps into workflows that run on triggers or schedules.

Cloud

–

Automation AnywhereAutomation AnywhereNo public result yet –

GumloopGumloop · $37/moNo public result yet –

IFTTTIFTTT · Free tier · $2.99/moNo public result yet –

LatenodeLatenode · Free tierNo public result yet –

MakeMake · Free tierNo public result yet –

Power AutomateMicrosoft · Free tier · $15/moNo public result yet –

Tray.aiTray.aiNo public result yet –

UiPathUiPath · Free tier · $25/moNo public result yet –

WorkatoWorkatoNo public result yet –

ZapierZapier · Free tier · $20/moNo public result yet

Local

–

n8nn8n · FreeNo public result yet –

AutoGPT PlatformSignificant Gravitas · FreeNo public result yet –

DifyLangGenius · FreeNo public result yet –

LangflowLangflow · FreeNo public result yet –

SimSim Studio · Free tier · $25/moNo public result yet –

ActivepiecesActivepieces · FreeNo public result yet –

Coze StudioByteDance (Coze) · FreeNo public result yet

Agent builders

Build AI agents that use your tools and act on your behalf, with little or no code.

Cloud

ChatGPTOpenAI · Free tier · $20/movia GPT-6 Astra$–

DustDust · Free tierNo public result yet –

AgentforceSalesforceNo public result yet –

Google Workspace StudioGoogleNo public result yet –

GumloopGumloop · $37/moNo public result yet –

LindyLindy · $30/moNo public result yet –

Microsoft Copilot StudioMicrosoftNo public result yet –

MindStudioMindStudio · Free tier · $20/moNo public result yet –

NotionNotion Labs · Free tier · $24/moNo public result yet –

Relevance AIRelevance AINo public result yet –

StackAIStackAINo public result yet –

TwinTwin · $29/moNo public result yet –

ZapierZapier · Free tier · $20/moNo public result yet

Local

–

AutoGPT PlatformSignificant Gravitas · FreeNo public result yet –

DifyLangGenius · FreeNo public result yet –

CrewAICrewAI · FreeNo public result yet –

SimSim Studio · Free tier · $25/moNo public result yet –

Coze StudioByteDance (Coze) · FreeNo public result yet –

KortixKortix · FreeNo public result yet

Quality vs. price

AutomationBench-AA (Artificial Analysis) score against API price (USD per 1M tokens, blended 3:1 input:output, log scale). Up and to the left is better; the line joins the models that nothing else beats on both.

Closed models
Open-weights models
Best-value frontier

Show as table

Model	Weights	AutomationBench-AA (Artificial Analysis)	Price	Frontier
Gemini 4 Argon	Closed	77.5%	$4/M tokens	Yes
Claude Sonnet 5.5	Closed	71.8%	$4/M tokens
Muse Wonder	Closed	71.3%	$8/M tokens
Claude Opus 5.5	Closed	69.5%	$8/M tokens
GPT-6 Astra	Closed	68.5%	$20/M tokens
DeepSeek-V4.1-Flash-1017	Open	66.6%	$0.52/M tokens	Yes
Grok 4.7	Closed	65.6%	$3/M tokens
GPT-6.1 Sol	Closed	64.9%	$4/M tokens
GLM-5.3	Open	62.2%	$2.15/M tokens
GLM-5.3-Flash	Open	60.4%	$0.24/M tokens	Yes
Gemini 3.8 Flash	Closed	59.9%	$1.5/M tokens
Mistral Large 4 (preview)	Closed	59.9%	$2.06/M tokens
Claude Fable 5.1	Closed	59.4%	$20/M tokens
MiMo-V2.6-Pro	Open	58.6%	$0.54/M tokens
Kimi K3	Open	58.3%	$6/M tokens
Qwen3.8-Max	Closed	56.2%	$3/M tokens
GPT-6 Luna	Closed	53.2%	$0.2/M tokens	Yes
Step 5 Preview	Open	51%	$1.43/M tokens
Muse Spark Open (816B-A42B)	Open	49.7%	$2/M tokens
Beam 1.0 (501B-A23B)	Open	48.5%	$1.5/M tokens
Qwen3.8-27B	Open	48.2%	$0.96/M tokens
Claude Haiku 5.5	Closed	35.4%	$0.2/M tokens
Nemotron 3.5 Super VL 120B-A12B	Open	14.8%	$0.18/M tokens	Yes
Inkling	Open	5%	$1.76/M tokens

Quality vs. setup

The same tools as the map, placed by how good the results can be (up) and how much setup they take (right). Top-left is the easiest way to top quality; further right buys control, privacy or price at the cost of setup. Models are left out: you reach them through these tools.

Cloud
Local: runs on your hardware
Hollow: tier inherited from the best model the tool offers

Not placed (no public result yet): n8n, AutoGPT Platform, Dify, Langflow, CrewAI, Sim, Activepieces, Coze Studio, Kortix, Dust, Agentforce, Automation Anywhere, Google Workspace Studio, Gumloop, IFTTT, Latenode, Lindy, Make, Microsoft Copilot Studio, MindStudio, Notion, Power Automate, Relevance AI, StackAI, Tray.ai, Twin, UiPath, Workato, Zapier

Everything, in order

All workflow automation options in quality order
Tier	Name	Approach	Runs	Rated by	Price	Skill
S	Gemini 4 Argon Google	Models	Cloud	77.5% on AutomationBench-AA (Artificial Analysis)	$4/M tokens	—
A	Claude Sonnet 5.5 Anthropic	Models	Cloud	71.8% on AutomationBench-AA (Artificial Analysis)	$4/M tokens	—
A	Muse Wonder Meta	Models	Cloud	71.3% on AutomationBench-AA (Artificial Analysis)	$8/M tokens	—
A	Claude Opus 5.5 Anthropic	Models	Cloud	69.5% on AutomationBench-AA (Artificial Analysis)	$8/M tokens	—
A	GPT-6 Astra OpenAI	Models	Cloud	68.5% on AutomationBench-AA (Artificial Analysis)	$20/M tokens	—
A	ChatGPT OpenAI	Agent builders	Cloud	via GPT-6 Astra	Free tier · $20/mo	beginner
B	DeepSeek-V4.1-Flash-1017 DeepSeek	Models	Local	66.6% on AutomationBench-AA (Artificial Analysis)	~320 GB VRAM	—
B	Grok 4.7 xAI	Models	Cloud	65.6% on AutomationBench-AA (Artificial Analysis)	$3/M tokens	—
B	GPT-6.1 Sol OpenAI	Models	Cloud	64.9% on AutomationBench-AA (Artificial Analysis)	$4/M tokens	—
B	GLM-5.3 Z.ai	Models	Local	62.2% on AutomationBench-AA (Artificial Analysis)	~425 GB VRAM	—
B	GLM-5.3-Flash Z.ai	Models	Local	60.4% on AutomationBench-AA (Artificial Analysis)	~185 GB VRAM	—
B	Gemini 3.8 Flash Google	Models	Cloud	59.9% on AutomationBench-AA (Artificial Analysis)	$1.5/M tokens	—
B	Mistral Large 4 (preview) Mistral AI	Models	Cloud	59.9% on AutomationBench-AA (Artificial Analysis)	$2.06/M tokens	—
C	Claude Fable 5.1 Anthropic	Models	Cloud	59.4% on AutomationBench-AA (Artificial Analysis)	$20/M tokens	—
C	MiMo-V2.6-Pro Xiaomi	Models	Local	58.6% on AutomationBench-AA (Artificial Analysis)	~585 GB VRAM	—
C	Kimi K3 Moonshot AI	Models	Local	58.3% on AutomationBench-AA (Artificial Analysis)	~1600 GB VRAM	—
C	Qwen3.8-Max Alibaba (Qwen)	Models	Cloud	56.2% on AutomationBench-AA (Artificial Analysis)	$3/M tokens	—
C	GPT-6 Luna OpenAI	Models	Cloud	53.2% on AutomationBench-AA (Artificial Analysis)	$0.2/M tokens	—
C	Step 5 Preview StepFun	Models	Local	51% on AutomationBench-AA (Artificial Analysis)	~346 GB VRAM	—
C	Muse Spark Open (816B-A42B) Meta	Models	Local	49.7% on AutomationBench-AA (Artificial Analysis)	~470 GB VRAM	—
C	Beam 1.0 (501B-A23B) Reflection AI	Models	Local	48.5% on AutomationBench-AA (Artificial Analysis)	~287 GB VRAM	—
C	Qwen3.8-27B Alibaba (Qwen)	Models	Local	48.2% on AutomationBench-AA (Artificial Analysis)	~17 GB VRAM	—
C	Claude Haiku 5.5 Anthropic	Models	Cloud	35.4% on AutomationBench-AA (Artificial Analysis)	$0.2/M tokens	—
C	Inkling Thinking Machines Lab	Models	Local	5% on AutomationBench-AA (Artificial Analysis)	~557 GB VRAM	—
C	K2-Horizon-375B-A23B MBZUAI Institute of Foundation Models	Models	Local	37.2% on AutomationBench-AA (Artificial Analysis)	~218 GB VRAM	—
C	Nemotron 3.5 Super VL 120B-A12B NVIDIA	Models	Local	14.8% on AutomationBench-AA (Artificial Analysis)	~72 GB VRAM	—
–	n8n n8n	Automation platforms	Local	No public result yet	Free	intermediate
–	AutoGPT Platform Significant Gravitas	Agent builders + Automation platforms	Local	No public result yet	Free	advanced
–	Dify LangGenius	Automation platforms + Agent builders	Local	No public result yet	Free	intermediate
–	Langflow Langflow	Automation platforms	Local	No public result yet	Free	intermediate
–	CrewAI CrewAI	Agent builders	Local	No public result yet	Free	advanced
–	Sim Sim Studio	Automation platforms + Agent builders	Local	No public result yet	Free tier · $25/mo	intermediate
–	Activepieces Activepieces	Automation platforms	Local	No public result yet	Free	intermediate
–	Coze Studio ByteDance (Coze)	Agent builders + Automation platforms	Local	No public result yet	Free	advanced
–	Kortix Kortix	Agent builders	Local	No public result yet	Free	advanced
–	Dust Dust	Agent builders	Cloud	No public result yet	Free tier	beginner
–	Agentforce Salesforce	Agent builders	Cloud	No public result yet	—	intermediate
–	Automation Anywhere Automation Anywhere	Automation platforms	Cloud	No public result yet	—	intermediate
–	Google Workspace Studio Google	Agent builders	Cloud	No public result yet	—	beginner
–	Gumloop Gumloop	Automation platforms + Agent builders	Cloud	No public result yet	$37/mo	intermediate
–	IFTTT IFTTT	Automation platforms	Cloud	No public result yet	Free tier · $2.99/mo	beginner
–	Latenode Latenode	Automation platforms	Cloud	No public result yet	Free tier	intermediate
–	Lindy Lindy	Agent builders	Cloud	No public result yet	$30/mo	beginner
–	Make Make	Automation platforms	Cloud	No public result yet	Free tier	intermediate
–	Microsoft Copilot Studio Microsoft	Agent builders	Cloud	No public result yet	—	intermediate
–	MindStudio MindStudio	Agent builders	Cloud	No public result yet	Free tier · $20/mo	beginner
–	Notion Notion Labs	Agent builders	Cloud	No public result yet	Free tier · $24/mo	beginner
–	Power Automate Microsoft	Automation platforms	Cloud	No public result yet	Free tier · $15/mo	intermediate
–	Relevance AI Relevance AI	Agent builders	Cloud	No public result yet	—	intermediate
–	StackAI StackAI	Agent builders	Cloud	No public result yet	—	beginner
–	Tray.ai Tray.ai	Automation platforms	Cloud	No public result yet	—	intermediate
–	Twin Twin	Agent builders	Cloud	No public result yet	$29/mo	beginner
–	UiPath UiPath	Automation platforms	Cloud	No public result yet	Free tier · $25/mo	intermediate
–	Workato Workato	Automation platforms	Cloud	No public result yet	—	intermediate
–	Zapier Zapier	Automation platforms + Agent builders	Cloud	No public result yet	Free tier · $20/mo	beginner

Benchmarks used

AutomationBench-AA (Artificial Analysis) by Artificial Analysis (benchmark developed by Zapier) (automated tests, Score: average share of each task's objectives completed across 657 business-workflow tasks in 40 simulated SaaS apps via REST APIs; a task with any guardrail (business-rule) violation scores zero) · snapshot 19 Oct 2026
32 of the 227 models on AA have a score (page: '32 of 227 models'); launched 2026-07-06 and since re-run on current versions (Gemini 4 Argon, Claude Opus/Sonnet 5.5, Fable 5.1, GPT-6 Astra/Sol/Luna, Grok 4.7, Muse Wonder, Kimi K3, GLM-5.3, DeepSeek V4.1 Flash 1017 and others; newest row Muse Wonder Full, released 2026-10-12). Run independently by AA on Zapier's private task subset, one run per task with a 50-turn cap; Zapier's own hosted board (zapier.com/benchmarks, v1.0.6, 126 rows, leader Gemini 4 Argon (High) 51.29%) uses a strict all-assertions-pass metric and is run by the benchmark's author. Models only: automation platforms and agent builders are not ranked, so tools could only inherit tiers via uses_models. Untracked: MiniMax-M3.1-Flash (72.52, #2).