Claude 5 Sonnet
currentClaude 5 Sonnet is Anthropic's lower-cost fifth-generation model and a strong default for coding, agents, writing, and professional work. It approaches Opus 4.8 on some agent tasks at higher effort, but its introductory pricing is temporary and its newer tokenizer can increase billed token counts.
Context window
1M tokens
Maximum output
128K tokens
Input price
$2 / 1M
Output price
$10 / 1M
Best for
- Everyday coding and agentic workflows
- Professional writing and analysis
- Computer-use and browsing agents
- Teams seeking a cheaper alternative to Opus 4.8
Not ideal for
- The hardest long-horizon work where Fable 5 is justified
- Pricing plans that assume the introductory rate is permanent
- Workflows that rely on non-default sampling parameters
- Native audio or video generation
PromptHQ 100-word challenge
We give every model the same miniature science-fiction mystery. The response is short enough to read in one chat bubble, while the premise tests creative writing, compression, foreshadowing, causal coherence, and instruction following.
Claude 5 Sonnet103 wordsWrite a science-fiction mystery in exactly 100 words.
A three-person research station on Europa receives a message from itself, timestamped 11 minutes in the future:
"Do not open the greenhouse door."
The door is opened anyway.
Plant two subtle clues. Give the warning a logical explanation without aliens, time travel, hallucinations, or simulations. The final sentence must change what the reader thought the warning meant.
Output only the story. Do not explain it.
The airlock cycled twice before the message came through—standard for a system clearing static, nothing more.
Voss read it aloud. Reyes laughed and opened the greenhouse door anyway, since ignoring warnings was the crew's running joke this rotation.
Nothing happened. The tomatoes hung heavy, undisturbed, oxygen levels nominal.
Eleven minutes later, the hull breach alarm sounded three corridors down, and Voss finally understood: the message hadn't warned them about the door. It had told them where to be standing when the real emergency hit—because opening it was the only thing that got both crewmates into the reinforced greenhouse before the wall gave way.
Run the same prompt yourself
Open Claude 5 Sonnet in Multichat with the evaluation prompt prefilled.
What the response revealed
The response contains 103 whitespace-delimited words, three over the required count.
It outputs only a story and avoids the prohibited explanations.
It never explains who sent the station's message or why its timestamp was eleven minutes ahead, and 'do not open' conflicts with the claimed intent to move the crew inside.
The reinforced greenhouse resolves the emergency, but the opening airlock detail does not clearly foreshadow that resolution.
The ending attempts a reinterpretation, but the warning's literal instruction works against the proposed meaning.
The prose is readable and compact, but the central logic does not survive scrutiny.
The model receives the prompt without web access or external tools. Claude 5 Sonnet is run through Claude CLI at medium reasoning effort. We preserve the response as generated apart from display rendering.
Performance
Claude 5 Sonnet benchmarks
Benchmark scores are sensitive to reasoning effort, harness, tools, token budget, prompt format, sampling, and evaluation date. Scores here retain their source and should not be treated as directly interchangeable unless the underlying setup matches.
Long-horizon agent tasks, pass@1
54.3%
multi-step agents · Anthropic3
Long-horizon agent tasks, pass@3
63%
multi-step agents · Anthropic3
Tool-assisted evaluation
81.6%
tool use · Anthropic3
Family position
Claude 5 positioning
Sonnet 5 is the value-oriented Claude 5 model. Anthropic positions it near Opus 4.8 on agentic work at substantially lower introductory token prices, while Fable 5 remains the maximum-capability option.
Claude Fable 5
Highest general capability
$10 input
$50 output
Claude Opus 4.8
Complex coding and enterprise
$5 input
$25 output
Claude Sonnet 5
Balanced speed and intelligence
$2 input
$10 output
API pricing
Per million text tokens
Input
$2
Cached input
$0.20
Cache write
$2.50
Output
$10
Where Sonnet 5 stands out
Built for demanding work
Everyday coding and agentic workflows are central to the model's positioning, rather than an incidental capability.
Long-context capacity
The published context window is 1,000,000 tokens, making the model a candidate for large documents, repositories, and sustained agent state.
Reasoning and tools
Reasoning is supported with low, medium, high, xhigh, max provider setting(s), and the model can participate in tool-using workflows through its available API surface.
Limitations to know
Benchmarks are configuration-sensitive
Scores can move substantially with the harness, tool access, effort setting, token budget, and evaluator. Treat the table as evidence, not a universal ranking.
Context size is not guaranteed recall
A large advertised window does not mean every detail is retrieved reliably at maximum length. Validate representative long-context workloads before deployment.
Product access differs from model capability
PromptHQ and gateway limits may expose fewer modalities, tools, or tokens than the provider's first-party API.
Capabilities and specifications
Knowledge cutoff
Jan 1, 2026
Inputs
text, image, PDF
Reasoning
Supported
Default effort
medium
Supported API features
Frequently Asked Questions
Sources
- 1What's new in Claude Sonnet 5
Anthropic · official documentation
- 2Introducing Claude Sonnet 5
Anthropic · official launch post
- 3Claude Sonnet 5 System Card
Anthropic · official system card
- 4PromptHQ model registry
PromptHQ · internal product configuration
Compare leading AI assistants
See how Claude 5 Sonnet handles your own work.
Try Sonnet 5 in MultichatAvailable on PromptHQ Plus, Max