LLM
Claude Sonnet 5 Review
Anthropic · LLM
Claude Sonnet 5 is Anthropic's most agentic model yet, delivering near-Opus 4.8 performance at significantly lower cost. It can autonomously use tools like
Claude Sonnet 5 is Anthropic's most agentic model yet, delivering near-Opus 4.8 performance at significantly lower cost. It can autonomously use tools like browsers and terminals, execute multi-step workflows, and maintain context across long agentic loops. Released alongside Fable 5 and Mythos 5 after the US government lifted national security restrictions.
Pros
- Fully agentic — uses browsers, terminals, tools autonomously
- Near-Opus performance at Sonnet price
- 200K+ context window
- Best-in-class instruction following
- Constitutional AI safety
Cons
- Requires careful prompt design for agentic loops
- Can be slow on long tool chains
- No native image generation
Use cases
- Autonomous coding agents
- Research workflows
- Multi-step document analysis
- Enterprise automation
- Long-form writing
Pricing
freemium — from $20/mo. A free tier is available.