The AI Model Landscape, October 2026: Every OpenAI, Claude and Gemini Model for Automation

The current AI models for Make, n8n and Zapier automations, as of 16 October 2026: OpenAI's lineup is GPT-6 Astra at the top ($10 input / $50 output per million tokens), GPT-6.1 Sol in the middle ($2 / $10) and GPT-6 Luna as the small model ($0.10 / $0.50). Anthropic's is Claude Fable 5.1 ($10 / $50), Claude Opus 5.5 ($4 / $20, the model Anthropic says to start with), Claude Sonnet 5.5 ($2 / $10) and Claude Haiku 5.5 ($0.10 / $0.50 for prompts up to 100,000 tokens). Google's is Gemini 4 Argon ($2 / $10 until 30 November 2026, then $4 / $20), Gemini 3.8 Flash ($0.75 / $3.75 until 31 December 2026) and the Flash-Lite models (3.5 at $0.30 / $2.50, 3.1 at $0.25 / $1.50). Make added the new OpenAI and Claude models to its AI modules between 4 September and 12 October 2026; Gemini 4 Argon was not in Make's release notes by 16 October, so it runs through the HTTP module for now.

This page is the catalog: what exists, what it costs and where it runs. Which model belongs on which workflow step is a separate decision, made in our routing guide; the month's platform changes are in automation platform updates. It is kept by Prem Patel, Make Level 5 Expert and Zapier Certified Expert at Nex Automations. Prices are US dollars per million tokens, standard tier, read from each vendor's own page on 16 October 2026. Batch processing is 50 percent off at all three vendors, per the OpenAI, Anthropic and Google pricing pages.

The lineup at a glance (16 October 2026)

Sources: the OpenAI, Anthropic and Google pricing pages and Make's release notes, read 16 October 2026. The tier column is our grouping from each vendor's own description. "No release note found" means no Make release note names the model, not that the module lacks it.

OpenAI models

OpenAI's models page says: "If you're not sure where to start, use GPT-6 Astra, our flagship model for complex reasoning and coding." It positions GPT-6.1 Sol as the balance of intelligence and cost and GPT-6 Luna for cost-sensitive, high-volume work. GPT-6 Sol (released 22 September) is still listed but was superseded a week later by GPT-6.1 Sol at the same $2 / $10 input and output price, so new builds should name gpt-6.1-sol. On every model below, a prompt over 272,000 input tokens is billed at 2x input and 1.5x output for the whole request. Release dates are from OpenAI's changelog.

GPT-6 Astra

• Released: 3 September 2026
• API ID: gpt-6-astra (model page)
• Price: $10 / $50; $20 / $75 over 272K input tokens
• Cached: $1.00
• Context: 1,050,000 tokens
• Max output: 128,000 tokens
• Good at in automation: the hardest single step, such as a research brief or a document built from many sources
• In Make: OpenAI modules from 11 September 2026, paid plans only in Simple text prompt (Make)
• Watch out: no custom temperature or top_p; tool calls need the Responses API

GPT-6.1 Sol

• Released: 29 September 2026
• API ID: gpt-6.1-sol (model page)
• Price: $2 / $10; $4 / $15 over 272K input tokens
• Cached: $0.10, 5 percent of input
• Context: 1,050,000 tokens
• Max output: 128,000 tokens
• Good at in automation: extraction, drafts and agent steps with tools; OpenAI calls it "near-Astra performance for complex work at a lower cost"
• In Make: Make AI Toolkit, OpenAI and Make AI Agent (New) apps from 9 October 2026 (Make); Create a Prompt module from 12 October, paid plans (Make)
• Watch out: tool calls need the Responses API; the none and minimal reasoning efforts are not supported

GPT-6 Luna

• Released: 22 September 2026
• API ID: gpt-6-luna (model page)
• Price: $0.10 / $0.50; $0.20 / $0.75 over 272K input tokens
• Cached: $0.01
• Context: 1,050,000 tokens
• Max output: 128,000 tokens
• Good at in automation: labels and routing at volume; Make suggests categorising support requests and form submissions
• In Make: Make AI Toolkit, OpenAI and Make AI Agent (New) apps from 25 September 2026 (Make)
• Watch out: in Chat Completions, function calling works only with reasoning effort set to none

GPT-5.6 Sol (now under Cyber models)

• Released: 9 July 2026
• API ID: gpt-5.6-sol (model page)
• Price: $4 / $20; $8 / $30 over 272K input tokens, a promotional price "available at least through November 21, 2026"
• Cached: $0.40
• Context: 1,050,000 tokens
• Max output: 128,000 tokens
• Good at in automation: scenarios already tuned on it; the pricing page now lists it under Cyber models
• In Make: GPT-5.6 support in the OpenAI modules from 17 July 2026 (Make)
• Watch out: twice GPT-6.1 Sol's price, with a promotion end date to watch

Anthropic Claude models

Anthropic's models overview says: "If you're unsure which model to use, start with Claude Opus 5.5 for most workloads." All four current models have a 1M-token context and 128K max output, and Fable 5.1, Opus 5.5 and Sonnet 5.5 bill the full window at standard rates (pricing). Two cautions cover all four: the tokenizer used since Claude 4.7 "produces approximately 30% more tokens for the same text" (pricing), and a non-default temperature, top_p or top_k returns a 400 error (deprecations). Dates are from each model page.

Claude Fable 5.1

• Released: 1 September 2026; retirement not sooner than 1 September 2027
• API ID: claude-fable-5-1 (model page)
• Price: $10 / $50
• Cached: $0.25 per cache read, 2.5 percent of input
• Context: 1M tokens
• Max output: 128K tokens
• Good at in automation: long-horizon agent runs where Opus 5.5 falls short; default effort high
• In Make: Make AI Agent (New), Make AI Toolkit, MCP Client and Anthropic Claude apps from 4 September 2026 (Make)
• Watch out: forced tool use returns an error

Claude Opus 5.5

• Released: 22 September 2026; retirement not sooner than 22 September 2027
• API ID: claude-opus-5-5 (model page)
• Price: $4 / $20
• Cached: $0.20 per cache read, 5 percent of input
• Context: 1M tokens
• Max output: 128K tokens (300K on the Batch API, beta)
• Good at in automation: agent steps that plan, call tools and recover; default effort medium
• In Make: Anthropic Claude app from 25 September 2026 (Make)
• Watch out: thinking cannot be disabled; on Make's Free plan, premium models such as Opus now run with Haiku 5.5 as the fallback (Make)

Claude Sonnet 5.5

• Released: 28 September 2026; retirement not sooner than 28 September 2027
• API ID: claude-sonnet-5-5 (model page)
• Price: $2 / $10
• Cached: $0.10 per cache read, 5 percent of input
• Context: 1M tokens
• Max output: 128K tokens (300K on the Batch API, beta)
• Good at in automation: long documents and detailed drafts a person approves
• In Make: Anthropic Claude app on paid plans from 2 October 2026 (Make); Make AI Toolkit and Make AI Agent (New) from 9 October (Make)
• Watch out: Make's Create a Prompt module limits tool choice to Auto with this model

Claude Haiku 5.5

• Released: 7 October 2026; retirement not sooner than 7 October 2027
• API ID: claude-haiku-5-5 (model page)
• Price: $0.10 / $0.50 for prompts up to 100,000 tokens; $0.50 / $2.50 above
• Cached: $0.01 per cache read ($0.05 above 100,000 tokens)
• Context: 1M tokens
• Max output: 128K tokens (300K on the Batch API, beta)
• Good at in automation: Anthropic names classification, routing, extraction and subagent tasks
• In Make: Create a Prompt module from 12 October 2026 (Make)
• Watch out: one prompt over 100,000 tokens, cache reads included, pays five times the rate for the whole request

Legacy Claude models still available

Sources: Claude pricing and model deprecations, read 16 October 2026.

Google Gemini models

Google's models page lists Gemini 4 Argon as its only Gemini 4 model. Release months below are the "latest update" each model page shows.

Gemini 4 Argon

• Released: generally available 13 October 2026 per Google's Gemini API changelog; unveiled 30 September, as reported by VentureBeat
• API ID: gemini-4-argon (model page)
• Price: $2 / $10 through 30 November 2026, $4 / $20 from 1 December
• Cached: $0.10, then $0.20; implicit caching only
• Context: 1,048,576 tokens, shared by input and output
• Max output: 65,536 by default, up to 1,048,576 with long decode
• Good at in automation: autonomous multi-step agents and very long outputs; Google calls it its "flagship reasoning model"
• In Make: not in the release notes by 16 October 2026; use the HTTP module
• Watch out: not on the free tier or the Google AI Pro plan

Gemini 3.8 Flash

• Released: September 2026
• API ID: gemini-3.8-flash (model page)
• Price: $0.75 / $3.75 through 31 December 2026, $1.50 / $7.50 from 1 January 2027
• Cached: $0.075, plus storage at $0.50 per million tokens per hour
• Context: 1,048,576 tokens
• Max output: 65,536 tokens
• Good at in automation: long multimodal inputs such as PDFs, audio and video; Google's "most intelligent Flash model"
• In Make: Google Gemini AI app from 4 September 2026 (Make)
• Watch out: the minimal thinking level returns an error

Gemini 3.5 Flash-Lite

• Released: July 2026
• API ID: gemini-3.5-flash-lite (model page)
• Price: $0.30 / $2.50
• Cached: $0.03, plus storage at $1.00 per million tokens per hour
• Context: 1,048,576 tokens
• Max output: 65,536 tokens
• Good at in automation: subagent tasks, document parsing and simple data extraction, per Google
• In Make: no release note names it; check the Google Gemini AI module's list or use the HTTP module
• Watch out: output costs over eight times input

Gemini 3.1 Flash-Lite

• Released: May 2026
• API ID: gemini-3.1-flash-lite (model page)
• Price: $0.25 / $1.50 ($0.50 input for audio)
• Cached: $0.025
• Context: 1,048,576 tokens
• Max output: 65,536 tokens
• Good at in automation: translation and transcription at scale, the jobs Google's guide leads with
• In Make: no release note names it
• Watch out: no computer use on this model

What changed since September 2026

The "before" prices are still listed on the OpenAI and Anthropic pricing pages, read 16 October 2026. Two shutdowns matter for older automations: OpenAI retires gpt-4-turbo, o1, o3-mini and o4-mini, among others, on 23 October 2026 and GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini on 14 January 2027 (OpenAI deprecations); Claude Sonnet 4.5 retires on 30 November 2026.

Where each model runs: Make, n8n and Zapier

Make. Models reach Make two ways: through Make's AI Provider, billed in credits, or through the vendor's app with your own API key. Since 9 October, Make AI Agents support prompt caching and a fallback connection that defaults to Make's AI Provider (Make). For EU teams, Make's data zones page says all OpenAI models in Make's AI Provider are processed in the US whatever data zone you select. Gemini 4 Argon goes through the HTTP module until a native module lands.

n8n. n8n's release notes do not name the current model versions, so we make no claim about any node's dropdown. They do record opt-in prompt caching (5-minute or 1-hour) on the Anthropic Chat Model node in n8n 2.37 (25 August 2026) and on the Anthropic node's Message operation in 2.42 (29 September 2026). Version 2.42 also renamed the Azure OpenAI Chat Model node to Azure AI Foundry Chat Model, since it covers that whole catalogue, Claude included. Any model with an HTTP API works from the HTTP Request node.

Zapier. Zapier's post of 6 October 2026, "Zapier Agents is now AI by Zapier", says an AI by Zapier step takes a Standard, Advanced or Premium model tier, or a named model. Default models are free; other tiers carry multipliers, so a step bills 1, 3 or 5 tasks, shown per run in Zap History. The post does not say which model sits in each tier.

Cheapest, largest context and best for agents

Drawn from the facts above, read 16 October 2026. Opinions are marked as our pick.

• Cheapest: GPT-6 Luna and Claude Haiku 5.5, tied at $0.10 / $0.50 with $0.01 cached input. Cheapest Gemini for new projects: 3.1 Flash-Lite at $0.25 / $1.50 (Gemini 2.5 Flash-Lite at $0.10 / $0.40 is still listed but limited to existing users).
• Most expensive: GPT-6 Astra and Claude Fable 5.1 at $10 / $50.
• Cheapest frontier model: Gemini 4 Argon at $2 / $10 until 30 November 2026; from 1 December it matches Claude Opus 5.5.
• Largest context: the GPT-6 models at 1,050,000 tokens, a hair above Gemini's 1,048,576 and Claude's 1M.
• Largest context at one price: Claude Fable 5.1, Opus 5.5 and Sonnet 5.5 across the full 1M window. Gemini 4 Argon also lists no long-context tier.
• Longest output: Gemini 4 Argon, up to 1,048,576 tokens with long decode.
• Best for agents, our pick: Claude Opus 5.5. Anthropic says to start there, it is built for long-running agentic work, cache reads cost 5 percent and it has been native in Make since 25 September.
• Best for agents on a budget, our pick: GPT-6.1 Sol, half Opus 5.5's price, with tools, multi-agent in beta and native Make modules since 9 October.
• Best for high-volume labels, our pick: GPT-6 Luna or Claude Haiku 5.5; the price is level, so your own test set decides.

How we choose

A catalog tells you what exists, not which model a step needs, and the answer is usually smaller than the newest name here. We score candidates on real items from the client's own workflow before anything goes live, then re-score when a vendor ships a new version. The method, the tier per job and the cost per 1,000 runs are in which AI model for which automation workflow.

Not sure which models your automations call?

If you do not know which models your scenarios, Zaps or workflows call today, or whether any sit on the October or January shutdown lists, an automation audit checks every AI step against this catalog. Or book a free 30-minute call and bring the account.

Frequently asked questions

What is the cheapest AI model for automation in October 2026?

GPT-6 Luna and Claude Haiku 5.5 tie at $0.10 per million input tokens and $0.50 per million output, with cached input at $0.01, per the OpenAI and Anthropic pricing pages read 16 October 2026. Haiku 5.5's rate covers prompts up to 100,000 tokens. Google's cheapest model for new projects is Gemini 3.1 Flash-Lite at $0.25 / $1.50.

What is the latest GPT model in October 2026?

GPT-6.1 Sol, released 29 September 2026 per OpenAI's API changelog read 16 October 2026, at $2 per million input tokens and $10 output. OpenAI's current lineup is GPT-6 Astra (3 September 2026, $10 / $50), GPT-6.1 Sol and GPT-6 Luna (22 September, $0.10 / $0.50). OpenAI's models page says to start with GPT-6 Astra when unsure.

What is the latest Gemini model in October 2026?

Gemini 4 Argon, generally available on 13 October 2026 per Google's Gemini API changelog, at $2 per million input tokens and $10 output until 30 November 2026 and $4 / $20 from 1 December, per Google's pricing page read 16 October 2026. Gemini 3.8 Flash ($0.75 / $3.75 until 31 December 2026) is the current mid model.

Which AI model has the largest context window in October 2026?

The GPT-6 models at 1,050,000 tokens, just ahead of Gemini at 1,048,576 and Claude at 1M, per the vendors' model pages read 16 October 2026. OpenAI bills a whole request at 2x input and 1.5x output past 272,000 input tokens, while Claude Fable 5.1, Opus 5.5 and Sonnet 5.5 keep one price across the full 1M (Anthropic pricing page). Gemini 4 Argon has the longest output, up to 1,048,576 tokens with long decode.

Which AI model is best for agents inside Make, n8n or Zapier?

Our pick on 16 October 2026 is Claude Opus 5.5 at $4 / $20 per million tokens: Anthropic's models overview says to start with it for most workloads, cache reads cost 5 percent of input and it has been in Make's Anthropic Claude app since 25 September 2026. The budget pick is GPT-6.1 Sol at $2 / $10, in Make since 9 October. This is our opinion from client builds, not a benchmark.

GPT-6.1 Sol or Claude Sonnet 5.5: what is the difference?

Same price: $2 per million input tokens, $10 output and $0.10 cached input, per the OpenAI and Anthropic pricing pages read 16 October 2026. GPT-6.1 Sol has a 1,050,000-token context and doubles its input price past 272,000 tokens; Claude Sonnet 5.5 has a 1M context at one price. Both are in Make's AI modules, Sonnet 5.5 since 2 October 2026 and GPT-6.1 Sol since 9 October.

Can I use Gemini 4 Argon in Make.com?

Not through a native module yet. Gemini 4 Argon became broadly available on 13 October 2026, as reported by VentureBeat, but was not in Make's release notes by 16 October 2026. Call the Gemini API from Make's HTTP module instead, with an API key on a paid plan: Google's pricing page says Argon is not on the free tier.

When do Gemini API prices go up?

Per Google's Gemini API pricing page read 16 October 2026, Gemini 4 Argon moves from $2 / $10 to $4 / $20 per million input / output tokens on 1 December 2026. Gemini 3.8 Flash moves from $0.75 / $3.75 to $1.50 / $7.50 on 1 January 2027. The two Flash-Lite models show no scheduled change.

What is the latest Claude model?

Claude Haiku 5.5, released 7 October 2026 per Anthropic's model page read 16 October 2026. The current lineup is Claude Fable 5.1 (1 September 2026), Opus 5.5 (22 September), Sonnet 5.5 (28 September) and Haiku 5.5. Anthropic's models overview says: "If you're unsure which model to use, start with Claude Opus 5.5 for most workloads."

Which older OpenAI and Claude models are being retired?

Per OpenAI's deprecations page read 16 October 2026, gpt-4-turbo, o1, o3-mini and o4-mini are among the models shutting down on 23 October 2026; GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini follow on 14 January 2027, with gpt-6-luna named as the replacement. Anthropic retires Claude Sonnet 4.5 on 30 November 2026. Any automation pinned to those IDs needs a new model first.

Related reading

• Which AI Model for Which Automation Workflow? (October 2026)
• Automation Platform Updates, October 2026
• The iPaaS Landscape in 2026
• Make.com AI Agents Guide
• How Much Does an AI Agent Cost to Build in 2026?
• Zapier vs Make vs n8n in 2026