---
title: Sonnet 5.5 handles everyday marketing work at half the cost of opus
description: Here is how event marketers can tell which model a job actually needs.
image: https://backstage.hive.co/hubfs/buildingwithsonnet55.png
---

[Skip to content](https://backstage.hive.co/backstage-articles/you-probably-dont-need-opus-for-most-of-your-work#main-content)

[![Backstage with Hive.co](https://backstage.hive.co/hs-fs/hubfs/bckwhive.png?width=170&height=47&name=bckwhive.png "Backstage with Hive.co")](https://backstage.hive.co/)

- [Tools](https://backstage.hive.co/tools)
- [Articles](https://backstage.hive.co/articles)
- [Community](https://backstage.hive.co/community)
- [Video](https://www.youtube.com/playlist?list=PLOTOWwzYFQr088c8WVspPzbPm4xX8RN2q)
- [Dispatch](https://backstage.hive.co/dispatch)

# Sonnet 5.5 handles everyday marketing work at half the cost of opus

Here is how event marketers can tell which model a job actually needs.

![](https://backstage.hive.co/hubfs/buildingwithsonnet55.png)

#### Published on Sep 29, 2026

#### [Written by Anthony Ramsay](https://backstage.hive.co/backstage-articles/author/anthony-ramsay)

 Category

#### [Backstage Community](https://backstage.hive.co/backstage-articles/tag/backstage-community)

The smartest AI model isn't always the one you should be using. If the work has a clear brief and you can tell whether the result is right, start with Sonnet 5.5, and save Opus for the work where judgment is the job.

Sonnet 5.5 costs $2 per million input tokens and $10 per million output tokens, while Opus 5.5 costs $4 and $20. Sonnet is also 30%+ faster than Sonnet 5, and Anthropic says it typically uses fewer tokens to finish the same work, which makes it up to 30% cheaper per task than its predecessor.

What matters more is what Anthropic is telling people to use it for. It points Sonnet 5.5 at well-scoped everyday work like repeated investigation, review, drafting, documents, spreadsheets, and fast iteration, which covers a lot of what most teams actually want AI doing every day.

## Stop defaulting to the biggest model

There's an understandable instinct with AI that says if one model is smarter, you should use that one. It doesn't hold up as a way to build AI into a business, and the better question is what's the cheapest, fastest model I can trust to do this job well.

If you're asking Claude to solve a messy strategic problem with incomplete information, reason across a lot of competing constraints, or work independently toward an ambiguous outcome, use Opus. Anthropic describes Opus 5.5 as the model for complex work requiring careful judgment and the hardest long-horizon problems.

Most work isn't that. Most work comes with a brief, and most work has an answer you can check.

## An event marketer's day is mostly checkable work

A normal day means researching an artist, drafting the presale copy, turning show notes into a campaign brief, comparing ticket sales to last week's, or summarizing a meeting into action items.

Those tasks still take intelligence, but they don't take frontier-level intelligence every time. Anthropic's own guidance says Sonnet 5.5 fits best when the task has a clear spec and a way to check the result, and that's a very useful rule for deciding which model should do the work.

![sonnet55\_workload](https://backstage.hive.co/hs-fs/hubfs/sonnet55_workload.png?width=1672&height=941&name=sonnet55_workload.png)

## This matters more when AI stops being something you ask

The cost difference between Sonnet and Opus is easy to ignore if you use Claude a few times a day. It gets much harder to ignore when Claude is working continuously.

Imagine an assistant that checks your business every morning. It looks at yesterday's ticket sales and compares them to the previous week, checks campaign performance, researches the artists you have going on sale, flags anything unusual, drafts the campaigns that need to go out, and updates a report. Then it does the same thing tomorrow, across another 50 shows.

At that point you care a lot about whether the model costs $2 or $4 per million input tokens, whether it finishes 30% faster, and whether it reaches the same useful result with fewer tokens. The interesting AI workflows happen over and over again, so every one of those differences gets multiplied.

![sonnet55\_pricing](https://backstage.hive.co/hs-fs/hubfs/sonnet55_pricing.png?width=1672&height=941&name=sonnet55_pricing.png)

## "Write me an email" is the least interesting version of this

A lot of teams still work with AI one task at a time, asking it to write this, summarize that, or give five subject lines. There's nothing wrong with that, but the bigger opportunity is handing AI a repeatable job.

Instead of "Write an on-sale email," try "Research the artist, look at the audience, review how similar campaigns have performed, draft the presale and on-sale campaigns, check them against our brand voice, watch ticket sales after the campaign goes out, and tell me if anything looks wrong."

The first saves a few minutes. The second removes a workflow, and once you want AI running workflows all day, the economics of the model start to matter a lot.

## Know when to move up to Opus

Opus still has a clear job, and Sonnet just changes when you need it. I'd reach for Opus when I'm asking Claude to:

- rethink the strategy for an underperforming tour
- synthesize conflicting signals across markets
- make a series of interconnected decisions with incomplete information
- work independently on a difficult problem for a long time
- make a call where the hard part is judgment rather than execution

Anthropic says essentially the same thing. Opus is the better choice for the hardest long-horizon work, while Sonnet is the better fit for clearly defined tasks with verifiable outcomes.

There's another useful signal here. Sonnet 5.5 lets developers increase how much reasoning it does through higher effort settings, but Anthropic specifically warns that once you push Sonnet to its highest effort levels, it gets slower and more expensive enough that you may lose the balance that makes Sonnet useful in the first place. At that point, consider Opus.

That's a pretty clean line. Don't force Sonnet to become Opus. Use it for the work it's good at, and escalate when the job actually requires more judgment.

## The everyday model is now good enough for serious work

Sonnet 5.5 gets remarkably close to Opus on some measures of real-world work. Anthropic says it lands just two points behind Opus 5.5 on GDPval-AA, an evaluation based on professional tasks across different occupations, and it scored 70.6% on Terminal-Bench 4.0, a test of agentic coding.

Those numbers are impressive, but the more important point is that the everyday model is becoming good enough for serious work. That means you can use AI more often, run it across more workflows, and give it more repeatable jobs, all without reaching for the most expensive model every time you want something done well.

The operating rule is getting pretty simple. If you have a clear brief and a checkable result, start with Sonnet. If the problem is messy, the outcome is ambiguous, and judgment matters, move to Opus.

For most teams, most of the time, that means starting with Sonnet 5.5.

Join the conversation with event marketers figuring this out together: [Backstage community](https://backstage.hive.co/community)

### [Anthony Ramsay](https://backstage.hive.co/backstage-articles/author/anthony-ramsay)

![hive-backstage](https://backstage.hive.co/hubfs/hive-backstage.png)

## SUBSCRIBE TO BACKSTAGE DISPATCH

Sharp takes and real strategies from live event marketers, straight to your inbox every week.

## Latest Articles

More articles

![Sonnet 5.5 handles everyday marketing work at half the cost of opus](https://backstage.hive.co/hubfs/buildingwithsonnet55.png)

[Backstage Community](https://backstage.hive.co/backstage-articles/tag/backstage-community)

September 29, 2026

#### Sonnet 5.5 handles everyday marketing work at half the cost of opus

Here is how event marketers can tell which model a job actually needs.

 Read more

![The owned list audit: how much of your audience can you reach without paying a platform](https://backstage.hive.co/hubfs/BackstagewithHive.png)

[Backstage Community](https://backstage.hive.co/backstage-articles/tag/backstage-community)

September 23, 2026

#### The owned list audit: how much of your audience can you reach without paying a platform

Followers, past buyers, email and SMS opt-ins are four different things. Here is a twenty-minute audit that tells you how many people you c...

 Read more

![How to Use AI to Make Sense of Your Ticketing Data](https://backstage.hive.co/hubfs/How%20to%20Use%20AI%20to%20Make%20Sense%20of%20Your%20Ticketing%20Data.png)

[Backstage Community](https://backstage.hive.co/backstage-articles/tag/backstage-community)

September 08, 2026

#### How to Use AI to Make Sense of Your Ticketing Data

Your ticketing export has been answering questions for years and nobody had time to ask. Here's how to get real answers out of it in an aft...

 Read more

[![Hive Black Logo svg](https://backstage.hive.co/hubfs/Hive%20Black%20Logo%20svg.svg "Hive Black Logo svg")](https://www.hive.co)

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Anthony Ramsay",
    "url" : "https://backstage.hive.co/backstage-articles/author/anthony-ramsay"
  },
  "dateModified" : "2026-09-29T16:43:28.868Z",
  "datePublished" : "2026-09-29T16:43:28.000Z",
  "headline" : "Sonnet 5.5 handles everyday marketing work at half the cost of opus",
  "image" : [ "https://backstage.hive.co/hubfs/buildingwithsonnet55.png" ],
  "mainEntityOfPage" : {
    "@id" : "https://backstage.hive.co/backstage-articles/you-probably-dont-need-opus-for-most-of-your-work",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://backstage.hive.co/hubfs/Hive%20logo_black.png"
    }
  }
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://www.hive.co/#organization",
  "@type" : "Organization",
  "description" : "Hive is an event marketing platform that helps promoters, venues, and festivals automate campaigns, engage fans, and sell more tickets.",
  "logo" : {
    "@type" : "ImageObject",
    "url" : "https://cdn.prod.website-files.com/60a44fa06f0feb314d1ad896/65aa7f0d8749d56b4f00d55a_FrameHive.svg"
  },
  "name" : "Hive",
  "sameAs" : [ "https://www.linkedin.com/company/hive-co" ],
  "url" : "https://www.hive.co/"
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://backstage.hive.co/#website",
  "@type" : "WebSite",
  "description" : "Tools and resources from Hive to help event professionals streamline marketing workflows and improve event promotion.",
  "inLanguage" : "en",
  "name" : "Hive Backstage",
  "publisher" : {
    "@id" : "https://www.hive.co/#organization"
  },
  "url" : "https://backstage.hive.co/"
}
```