Anthropic · tested in Taskade Genesis
Claude Sonnet 5 in Taskade
Claude Sonnet 5 from Anthropic runs on Pro and up, with your own API key. We last tested it on Oct 18, 2026 in TSK-1, our app-building test.
- Status
- Own key
- Plan
- Pro and up, with your own API key (BYOK)
- Where to pick it
- Agent settings
- Your own key
- Anthropic
- Last test
- Context window
- 1Mtokens
- Max output
- 128Ktokens
- Released
- Knowledge cutoff
- Where it runs
- 8services
TSK-1 scorecard for the Claude family
TSK Score
Claude family, across real apps made in Taskade Genesis
- Interfaceready to share
- Strong
- Taskcompletes your task
- Strong
- Memorykeeps your data
- Emerging
- Adapthandles follow-up edits
- Leading
Our most carefully finished result in August.
MoreLess
In September, one dashboard could not read its own sample data and a longer build stopped before the app was written. In October its reading list and its family intake forms saved every entry, and its teacher page showed each answer choice in plain words.
Published scores
Benchmark scores that Anthropic and test sites published for Claude Sonnet 5. Taskade did not re-run them. Lab-reported marks a score from Anthropic itself.
| Benchmark | Score | Source |
|---|---|---|
| SWE-Bench Verifiedresolved · AnthropicLab-reported | 85.2% | AnthropicLab-reported |
| SWE-Bench Proresolve rate · AnthropicLab-reported | 63.2% | AnthropicLab-reported |
| Terminal-Benchsuccess rate · v2.1 · Terminus-2 · AnthropicLab-reported | 80.4% | AnthropicLab-reported |
| OSWorld-Verifiedsuccess rate · AnthropicLab-reported | 81.2% | AnthropicLab-reported |
| BrowseCompaccuracy · single agent · AnthropicLab-reported | 84.7% | AnthropicLab-reported |
More published scores (2)
LMArena scores
| Benchmark | Score | Source |
|---|---|---|
| Text arena±4 · 49,595 votes · Rank 58 of 416 · LMArena | 1,461 | LMArena |
| WebDev arena±6 · 15,665 votes · Rank 45 of 145 · LMArena | 1,541 | LMArena |
Scores from the LMArena leaderboard dataset (CC BY 4.0). People vote between two answers, and the votes set the score. Taskade picked the Overall category and the default setting of each model.
Scores collected by models.dev (MIT). Each source links the page that published the score.
Versions tested
Open a version to read its findings. We ran these tests with an API key from the lab.
Claude Sonnet 5Tested Oct 202614 findings
Our most carefully finished result in August. In September, one dashboard could not read its own sample data and a longer build stopped before the app was written. In October its reading list and its family intake forms saved every entry, and its teacher page showed each answer choice in plain words.
Opened its own app and checked the assistant's answers: the only model to verify its work this way.
Most carefully finished result, with eight minor items left.
Built a contacts database instead of the requested sign-up form after two stalls.
Four attempts stalled before the sign-up form was completed.
Stopped after asking its questions on the client sign-up form, and wrote the plan instead of the app.
Typed an 81-row scoring rubric and a 36-field sign-up table, then stopped at an approval prompt before writing the app.
Built a three-page match tracker whose dashboard reported zero matches over the eight matches it had entered itself.
Built the family intake form so it saved every answer and both photos, then added the new teacher question and put the change live.
Built the family intake form twice. Both builds saved every answer and both photos and took the new teacher question, and both showed sample families on the teacher page with no label.
Built the Dota 2 match tracker twice, each with a win-rate trend on the dashboard, and both saved a logged match. One ran wider than a phone screen. In the other the change added a second heroes page beside the one it had already built.
Built a reading list that saved a book added from the app, and each status filter showed the right books.
Built the family intake form so it saved every answer and both photos, and its teacher page showed each answer choice in plain words. Each saved family had a blank name in the project list.
Added the new teacher question, and the answer saved. Its closing summary described the teacher page as private-feeling, but anyone with the link could open it.
Added an Our Families page with sibling and family-activity fields to an intake app that another model had built. It locked the page with a class password written into the code of the app, and its closing summary said that this was not real security.
Results by task
Model facts
The token limits are the most the model accepts. Taskade can run it with lower limits.
- Last updated
- Input types
- Text, Images, PDF
- Output types
- Text
- Reasoning
- Yes
- Tool calling
- Yes
- File attachments
- Yes
- Structured output
- Yes
- Open weights
- No
- Reasoning effort
- LowMediumHighExtra highMax
Published specs from the model maker, via models.dev (MIT).
Run Claude Sonnet 5 with your own key
Pick it with your own key in a custom agent.
| Key provider | Plan |
|---|---|
| Anthropic | Pro and up |
Where it runs
Services outside Taskade that serve Claude Sonnet 5, from models.dev. Each link opens in a new tab. In Taskade, your own key works only through the key providers on this page.
How to choose Claude Sonnet 5
- Open the agent settings.
- Under AI Model, choose the model.
- Under Bring Your Own Key, add or select your provider key.
Where Claude ranks
TSK Score out of 100.
Compare Claude
Other Anthropic models
FAQ
Which Taskade plans include Claude Sonnet 5?
Claude Sonnet 5 runs on Pro and up, with your own API key. Your provider bills the model use to your own account.
Has Taskade tested Claude Sonnet 5?
Yes. We last tested Claude Sonnet 5 on Oct 18, 2026 in the TSK-1 Benchmark. Our most carefully finished result in August. In September, one dashboard could not read its own sample data and a longer build stopped before the app was written. In October its reading list and its family intake forms saved every entry, and its teacher page showed each answer choice in plain words.
How do I choose Claude Sonnet 5 in Taskade?
Open the agent settings. Under AI Model, choose Claude Sonnet 5. Under Bring Your Own Key, select your provider key.
Can I use my own API key for Claude Sonnet 5?
Yes. In a custom agent, you can pick Claude Sonnet 5 with your own key from Anthropic (Pro and up). Your provider bills the model use to your own account.
Who makes Claude Sonnet 5?
Anthropic makes Claude Sonnet 5. Taskade runs it on Pro and up, with your own API key, beside frontier models from other top AI labs.