Customer testimonials published by Anthropic for Claude Sonnet 5.5

Anthropic News · Veröffentlicht am 28. Sept. 2026

Anthropic’s Claude Sonnet 5.5 announcement includes named customer testimonials from Epic Games, Base44, Slack and Zendesk. These customers describe their own engineering, app-building, assistant and support tests. The accounts are selected and published by Anthropic; their reported results have not been independently verified by nafyi. Lies 4 Standpunkte mit Belegen und Links zu den Originalquellen.

Daniel Vogel Gabriel Grinberg Curtis Allen Abhinay Kathuria

Kernaussagen lesen Herausgeber öffnen

Auf einen Blick

Epic reports higher-tier quality on engineering tasks
In a customer testimonial published by Anthropic, Epic Games’ Daniel Vogel says Sonnet 5.5 met the quality bar he expected from a higher-tier model in early system-design and data-flow reviews, while handling long engineering tasks with less prescriptive prompting.
Unterstützendes Moment lesen · Source text block 30 · customer quotation
Base44 reports fewer iterations across 118 app builds
In a customer testimonial published by Anthropic, Base44’s Gabriel Grinberg reports that Sonnet 5.5 matched Opus 5’s scores across 118 app builds, averaging 3.6 iterations per build versus 7.7 for Opus 5.
Unterstützendes Moment lesen · Source text block 34 · customer quotation
Slack reports better offline results with fewer tokens
In a customer testimonial published by Anthropic, Slack’s Curtis Allen reports that Sonnet 5.5 outperformed Sonnet 5 on almost all of their offline Slackbot evaluations without prompt changes, using fewer steps and about 14% fewer output tokens.
Unterstützendes Moment lesen · Source text block 41 · customer quotation
Zendesk reports fewer wrong decisions and faster processing
In a customer testimonial published by Anthropic, Zendesk’s Abhinay Kathuria says Sonnet 5.5 made fewer wrong decisions on hundreds of support use cases and processed tickets 20% faster than the Claude models Zendesk was using at the time.
Unterstützendes Moment lesen · Source text block 42 · customer quotation

Wichtige Passagen4

Zugeordnete Passagen mit dem Kontext zur Überprüfung. Öffnen Sie den Originaltext, um die Quelle zu prüfen.

Source text block 34 · customer quotation

Application building iterations

Base44 reports fewer iterations across 118 app builds

Originalauszug

“Across 118 real app builds, Claude Sonnet 5.5 produced apps that scored level with Opus 5. It got there in 3.6 iterations per build on average, where Opus 5 took 7.7. It had the fewest failed tool calls of any model we compared. It also rarely stopped mid-build to ask the user a question, so fewer builds stall waiting on someone to answer.”

Gabriel Grinberg·28. Sept. 2026

Source text block 41 · customer quotation

Assistant evaluation efficiency

Slack reports better offline results with fewer tokens

Originalauszug

“Without changing any of our prompts, Claude Sonnet 5.5 did better than Sonnet 5 on almost all of our offline Slackbot evals, in fewer steps and with about 14% fewer output tokens. When someone gives Slackbot a task, quality and speed are what matter most, and Sonnet 5.5 allows Slackbot to deliver better outcomes for users, faster.”

Curtis Allen·28. Sept. 2026

Source text block 42 · customer quotation

Customer support workflows

Zendesk reports fewer wrong decisions and faster processing

Originalauszug

“We fed Claude Sonnet 5.5 hundreds of real support use cases across replies and escalation requests. It made fewer wrong decisions and resolved tickets faster than the Claude models we use in production today. Tickets were processed 20% faster, getting our customers the help they need without the wait.”

Abhinay Kathuria·28. Sept. 2026

Source text block 30 · customer quotation

Coding task quality

Epic reports higher-tier quality on engineering tasks

Originalauszug

“In Epic’s early testing, Claude Sonnet 5.5 cleared the same quality bar you’d expect from a higher-tier model, holding up on a system design audit and a data flow review. The new model managed tens of thousands of lines of code for gameplay system architecture, kept responses snappy, handled multi-hour tasks, and delivered with less prescriptive prompting.”

Daniel Vogel·28. Sept. 2026

Quelle & Methodik

Diese Standpunkte sind mit ihren Originalquellen verknüpft. Paraphrasen sind gekennzeichnet und keine wörtlichen Zitate.

Transkript oder Quellenmaterial öffnen Ein Problem melden