Claude Sonnet 4
Claude Sonnet 4 by Anthropic is a hybrid reasoning model with extended thinking + tool use, parallel tool calls, enhanced steerability, and free-user access.
Model Overview
Capabilities, design details, and architectural traits
Claude Sonnet 4 - Frontier Hybrid Reasoning for High-Volume
Claude Sonnet 4 is a hybrid reasoning large language model from Anthropic that operates in two documented modes: a standard mode for near-instant responses and an extended thinking mode for deeper reasoning. It is the first Claude 4 generation model made available to free-tier users, positioning it as the practical access point for the Claude 4 capability tier.
| Trait | Detail |
|---|---|
| Dual-mode architecture | Switches between standard and extended thinking modes per task |
| Extended thinking with tool use | Alternates between reasoning and external tools (e.g. web search) within a single extended thinking session |
| Parallel tool execution | Calls multiple tools simultaneously rather than sequentially |
| Enhanced steerability | Documented as more precisely responsive to operator and user instructions than its predecessor |
| Shortcut and loophole reduction | 65% less likely than Sonnet 3.7 to use shortcuts or loopholes on agentic tasks |
| Thinking summarization | Uses a smaller secondary model to condense long thought chains; raw chains available via opt-in Developer Mode |
| ASL-2 safety classification | Released under Anthropic's AI Safety Level 2 Standard, the same tier as its predecessor Sonnet 3.7 |
| Free-user availability | Available to Anthropic's free-plan users, unlike Claude Opus 4 which requires a paid plan |
Memory When Given File Access
When developers provide Claude Sonnet 4 with local file access, it can extract and save key facts to memory files, enabling continuity across longer sessions. This capability is shared with Claude Opus 4 in the same generation and is not built into the base model - it requires the developer to supply file access in the application layer.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | Anthropic |
| Release Date | May 22, 2025 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Jan 2025 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1M | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $3 | $4 | $2 Best Input Pricing |
| Output Pricing | $15 | $20 | $10 Best Output Pricing |
| Modalities | |||
| Inputs | imagetextfile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 16.6 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Anthropic
Other models by Anthropic
Claude Opus 5.5
Claude Sonnet 5.5
Claude Fable 5.1
Claude Opus 5
Top AI Models
Leading alternatives by intelligence score
