Arsaze
Let Claude, ChatGPT or Gemini edit your real timeline - an AI-native video editor for agents…
Free+Paid
- Web app
- Windows
What is Claude Sonnet 5.5?
The second model in the Claude 5.5 family: 30%+ faster than Sonnet 5 at up to 30% lower cost per task. Claude Sonnet 5.5: the everyday half of the 5.5 family Claude Sonnet 5.5 is Anthropic’s second Claude 5.5 model, announced September 28, 2026.
Full review
Claude Sonnet 5.5: the everyday half of the 5.5 family
Claude Sonnet 5.5 is Anthropic’s second Claude 5.5 model, announced September 28, 2026. Anthropic describes it as a clear upgrade over Sonnet 5: output 30%+ faster and up to 30% lower cost per task, at the same list price of USD 2 per million input and USD 10 per million output tokens. It complements Opus 5.5, which Anthropic still positions as the stronger choice for complex, judgment-heavy work.
Vendor-reported gains, especially in coding
The headline numbers are Anthropic’s own: 70.6% on Terminal-Bench 4.0 against Sonnet 5’s 10.3%, 55.5% on CursorBench 4.0, and 1844 on GDPval-AA v2.1, two points below Opus 5.5. Anthropic also cites strong long-horizon and image-understanding results, including a first for a Sonnet model: beating Pokemon Red from screenshots alone. Early tester quotes from Epic Games, SpaceXAI, Slack, Zendesk, Lovable, Atlassian and Unity describe fewer tool calls, faster
Claude Sonnet 5.5: the everyday half of the 5.5 family
Claude Sonnet 5.5 is Anthropic’s second Claude 5.5 model, announced September 28, 2026. Anthropic describes it as a clear upgrade over Sonnet 5: output 30%+ faster and up to 30% lower cost per task, at the same list price of USD 2 per million input and USD 10 per million output tokens. It complements Opus 5.5, which Anthropic still positions as the stronger choice for complex, judgment-heavy work.
Vendor-reported gains, especially in coding
The headline numbers are Anthropic’s own: 70.6% on Terminal-Bench 4.0 against Sonnet 5’s 10.3%, 55.5% on CursorBench 4.0, and 1844 on GDPval-AA v2.1, two points below Opus 5.5. Anthropic also cites strong long-horizon and image-understanding results, including a first for a Sonnet model: beating Pokemon Red from screenshots alone. Early tester quotes from Epic Games, SpaceXAI, Slack, Zendesk, Lovable, Atlassian and Unity describe fewer tool calls, faster ticket handling and shorter build loops. These are vendor-run evaluations and selected testimonials; independent verification is still catching up.
Cost, effort and where it runs
List pricing matches Sonnet 5, with USD 0.20 cache reads and USD 2.50 cache writes per million tokens. Anthropic says Sonnet 5.5 needs far fewer tokens per task, which is where the claimed savings come from. Adjustable effort levels trade cost and speed against quality: Medium by default in the Claude apps and Claude Code, High on the Claude Platform. The model is available in the Claude apps on web, iOS, Android and desktop, on the Claude Platform as claude-sonnet-5-5, and on AWS, Google Cloud and Microsoft Azure, with zero data retention availability.
Migration and safeguards
One breaking change matters: if you run Sonnet with thinking off, Anthropic requires the new between_tools setting before moving to Sonnet 5.5. On safety, Anthropic reports improved or matching alignment on its roughly 1,850-scenario behavioral audit, and Sonnet 5.5 is the first Sonnet to ship with cyber safeguards plus reasoning-extraction classifiers; biology safeguards match Sonnet 5. Claude Haiku 5.5 is announced for the coming weeks but not yet available.
Ashley Richmond’s editorial view
On paper this is the new default Claude for most everyday work: much faster than Sonnet 5, cheaper per task by Anthropic’s accounting, and near Opus 5.5 on several knowledge-work measures. I would still treat the benchmarks as vendor-reported until third-party results accumulate, and keep Opus 5.5 for the hardest open-ended problems. Provisional desk-research score: 4.2/5.
Desk-research disclosure: Official announcement, pricing, documentation and legal pages reviewed September 30, 2026. No hands-on model testing, API calls or independent benchmarking. All performance and cost-per-task figures are Anthropic-reported. Founders, founding year, headquarters and team size remain blank.
Who it's for
Best for
- Developers and teams who want a fast
- lower-cost model for everyday coding
- documents and agentic tasks across apps
- API and major clouds.
Not ideal for
- Buyers needing independently verified benchmark results
- the most complex open-ended reasoning (Opus 5.5)
- or high-volume cost-sensitive workloads (Haiku 5.5 is not out yet).
How we tested
Ashley Richmond desk research September 30, 2026. Official announcement, pricing page, models documentation, system card reference and legal pages reviewed; no hands-on model testing, API calls or account benchmarking. All performance, speed, cost-per-task and alignment figures are vendor-reported from Anthropic pages and attributed as such; footnoted caveats (pre-release evaluation bug, unreleased Haiku 5.5) included. Company name Anthropic from the official site; founders, founding year, headquarters and team size left blank. Media: logo is the official Claude starburst rasterized at high resolution from the claude.com header SVG lockup, full extent with white padding (no screenshot crop). claude.ai requires sign-in, so no anonymous web chat UI was capturable; gallery mixes one official web UI still (composer box from support.claude.com help center) with official App Store vendor-submitted iOS app UI stills (hero and second gallery), disclosed here as iOS app UI, not web UI. Sources: https://www.anthropic.com/claude-sonnet-5-5 ; https://www.anthropic.com/pricing ; https://docs.anthropic.com/en/docs/about-claude/models ; https://www.anthropic.com/legal/privacy ; https://www.anthropic.com/legal/consumer-terms
At a glance
| Pricing | Free+Paid |
|---|---|
| Free plan | Yes |
| Free trial | No |
| Pricing page | View pricing |
| Platforms | Web app, iOS (iPhone & iPad), Android, macOS, Windows, API / SDK |
| API | No |
| Company | Anthropic |
| Apps Insight Score | 46/100 |
| Implementation | Medium |
| Learning curve | Intermediate |
Why this app scored 46/100
Editorial Review 27/30
- Product quality 9/10 9/10
- Ease of use 4/5 4/5
- Feature depth 5/5 5/5
- Innovation 4/5 4/5
- Value for money 3/3 3/3
- Recommendation confidence 2/2 2/2
Trust & Verification 6/25
- AppsInsight badge installed Badge not installed 0/15
- HTTPS website HTTPS 1/1
- Business support email Missing 0/1
- Privacy policy published Provided 2/2
- Terms of service published Provided 1/1
- Security certifications None listed 0/2
- Knowledge base / help center Provided 2/2
- Social profiles linked 0 profiles linked 0/1
Community 0/20
- Verified reviews & rating 0 reviews (5 needed) 0/15
- Recent reviews No reviews yet 0/5
Product Profile 10/20
- Detailed description 402 words 3/3
- Screenshots 2 screenshots 1/3
- Demo video No demo video 0/2
- FAQ 6 items 1/2
- Feature list 8 items 1/2
- Pricing published Plan tiers published 3/3
- Pros & cons Pros & cons listed 1/1
- Integrations 2 integrations 0/2
- API available No 0/1
- Company details 1 of 3 details filled 0/1
Freshness 3/5
- Listing recently updated Updated 0 days ago 3/3
- Recent changelog entry No changelog entries 0/2
Editorial points are assigned by AppsInsight editors and cannot be purchased or influenced.
Integrations
Built for
Industry
SaaS & technology
How Claude Sonnet 5.5 compares
| Feature | Claude Sonnet 5.5 | Arsaze | reap | INBOT Studio |
|---|---|---|---|---|
| Rating | 4.2 / 5 | 4.2 / 5 | 4.0 / 5 | 3.8 / 5 |
| Pricing | Free+Paid | Free+Paid | Free+Paid | Free+Paid |
| Starting price | — | — | $9.99/mo billed annually | $14.90/mo; annual view from $11.92/mo |
| Free plan | Included | Included | Included | Included |
| Free trial | No | No | No | No |
| Platforms | Web app, iOS (iPhone & iPad), Android, macOS, Windows, API / SDK | Web app, Windows | Web app | Web app |
| Best for | Developers and teams who want a fast, lower-cost model for everyday coding, documents and agentic tasks across apps, API and major clouds. | Video creators and editors who want a real NLE their AI assistant can drive - text-based editing, grading, client review and generation on one timeline, with free local exports. | Podcasters and video creators turning long recordings into clips Global teams needing captions and dubbed versions Developers automating video repurposing through API or MCP Social teams publishing video-first content | Interior designers, freelancers and small studios preparing concepts and client proposals; teams that want to refine materials and export draft models to CAD or SketchUp. |
Claude Sonnet 5.5 vs Arsaze Claude Sonnet 5.5 vs reap Claude Sonnet 5.5 vs INBOT Studio
Key features
Faster generation
Anthropic reports output 30%+ faster than Claude Sonnet 5, its fastest Sonnet to date; noticeable in interactive coding and writing loops.
Lower cost per task
List price matches Sonnet 5 (USD 2 per million input, USD 10 per million output tokens, USD 0.20 cache reads), but Anthropic says it needs far fewer tokens per task, costing up to 30% less per task.
Agentic coding
Vendor-reported 70.6% on Terminal-Bench 4.0 (Sonnet 5: 10.3%), 55.5% on CursorBench 4.0 and 46.2-52.1% on FrontierCode 1.1; early testers report fewer tool calls per build.
Knowledge work and documents
1844 on GDPval-AA v2.1, two points below Opus 5.5 per Anthropic; positioned for polished documents, slides and spreadsheets with a noted eye for design.
Effort levels
Adjustable effort balances cost, speed and quality; Medium is the default in Claude apps and Claude Code, High on the Claude Platform.
Vision and computer use
Vendor-reported 80.1% partial on OSWorld 2.1 and 61.6% on Chartography without tools; first Sonnet to beat Pokemon Red working only from screenshots, per Anthropic.
Broad availability
In Claude apps on web, iOS, Android and desktop; Claude Platform API as claude-sonnet-5-5; Amazon Web Services, Google Cloud and Microsoft Azure; zero data retention availability.
Safeguards and alignment
First Sonnet with cyber safeguards and reasoning-extraction classifiers; biology safeguards match Sonnet 5; Anthropic reports improved or matching alignment on its ~1,850-scenario behavioral audit.
Key benefits
- Handle well-scoped everyday coding and writing tasks quickly at lower cost.
- Iterate fast at Medium effort, then raise effort for harder problems.
- Use one model across apps, API and the major clouds.
- Keep Opus 5.5 reserved for complex, judgment-heavy work.
Pricing
Free+Paid
Free
0 /mo
- USD 0. Chat on web
- iOS
- Android and desktop; code generation and data visualization; web search; memory; file creation and code execution; connectors with remote MCP; extended thinking. Usage limits apply.
Pro
20 /mo
- USD 20/month billed monthly
- or USD 17/month with annual subscription (USD 200 up front). More usage
- Claude Code
- Claude Cowork
- Claude Design
- Claude Science
- unlimited projects
- Research
- more models
- Claude for Microsoft 365.
Max
100 /mo
- From USD 100/month. 5x or 20x more usage than Pro
- higher output limits
- early access to advanced features
- priority access at high traffic times.
API (pay-as-you-go)
- Usage-priced: USD 2 per million input tokens
- USD 10 per million output tokens
- USD 0.20 cache reads
- USD 2.50 cache writes. Same list price as Sonnet 5; Batch API 50% off per Anthropic docs. Model ID claude-sonnet-5-5.
Pros & Cons
Pros
- Large measured jump over Sonnet 5 on vendor-run coding and knowledge-work benchmarks.
- Same list price as Sonnet 5 with lower reported cost per task.
- Available across apps, API, AWS, Google Cloud and Azure at launch.
- Detailed public announcement, system card and migration notes.
Cons
- Benchmarks are vendor-run; independent verification is still limited days after launch.
- Anthropic itself positions Opus 5.5 as stronger for complex, open-ended work.
- Thinking-off users face a breaking migration to the between_tools setting.
- Claude Haiku 5.5 is announced but not yet available, and plan usage limits apply.
Screenshots & media
Claude Sonnet 5.5 alternatives
View all alternatives
reap
AI clipping, captions, dubbing, and social publishing for long-form video.
Free planPaid from $9.99/mo billed annually
- Web app
INBOT Studio
Turn plans, room photos and sketches into editable 3D models, AI renders and targeted material edits.
Free planPaid from $14.90/mo; annual view from $11.92/mo
- Web app
Munch Studio
AI that plans, creates, schedules, and tracks your small business social media.
Paid Subscription$38/mo billed annually7-day trial
- Web app
CopyCopter
AI-assisted video creation and editing with direct social sharing
Usage-basedPay-as-you-go credits (~$0.10/credit)
- Web app
More Generative AI apps
INBOT Studio
Turn plans, room photos and sketches into editable 3D models, AI renders and targeted material edits.
Free planPaid from $14.90/mo; annual view from $11.92/mo
- Web app
Arsaze
Let Claude, ChatGPT or Gemini edit your real timeline - an AI-native video editor for agents…
Free+Paid
- Web app
- Windows
Munch Studio
AI that plans, creates, schedules, and tracks your small business social media.
Paid Subscription$38/mo billed annually7-day trial
- Web app
User reviews
No reviews yet. Be the first to review this app.
Write a review
Already have an account? Log in to track your reviews.
Frequently asked questions
Is Claude Sonnet 5.5 free to use?
Claude has a free plan on web, iOS, Android and desktop with usage limits. Paid Pro (USD 20 monthly or USD 17/month annual) and Max (from USD 100/month) raise limits, and the API bills per token. Prices from the official pricing page; tax not included.
How do developers access it?
On the Claude Platform with model ID claude-sonnet-5-5, plus Amazon Web Services, Google Cloud and Microsoft Azure. Anthropic says it is available with zero data retention, as with Opus 5.5 and Sonnet 5.
What does the API cost?
USD 2 per million input tokens and USD 10 per million output tokens, USD 0.20 for cache reads and USD 2.50 for cache writes, per the official announcement. Anthropic says fewer tokens per task make it up to 30% cheaper per task than Sonnet 5.
Sonnet 5.5 or Opus 5.5?
Anthropic positions Sonnet 5.5 for well-scoped everyday tasks, bug fixes and polished documents, and Opus 5.5 for complex work requiring careful judgment. Sonnet 5.5 is faster and cheaper per task; Opus 5.5 remains stronger on open-ended work.
What breaks when migrating from Sonnet 5?
If you run Sonnet with thinking off, Anthropic requires switching to the new between_tools setting, which keeps up-front thinking off, before moving to Sonnet 5.5. See the official migration guide.
Were the benchmarks independently verified?
Most figures come from Anthropic's own announcement and system card. Footnotes note Artificial Analysis ran GDPval-AA and AA-Briefcase on a pre-release deployment with a structured-output bug, since fixed. Treat scores as vendor-reported until more third-party results land.