Anthropic ships Claude Sonnet 5 at $2 per million tokens
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic released Claude Sonnet 5, positioning it as a midsize model that plans, uses browsers and terminals, and runs autonomously — capabilities that "just a few months ago" required larger and more expensive models
- Sonnet 5 launches at $2 per million input tokens and $10 per million output tokens through August 31, after which pricing jumps to $3 per million input and $15 per million output
- The new model is cheaper than Anthropic's Opus 4.8, OpenAI's GPT-5.5, and Google's Gemini 3.1 Pro, but more expensive than Google's Gemini 3.5 Flash
- Sonnet 5 scores 63.2% on an agentic coding benchmark — below Opus 4.8's 69.2% but well above predecessor Sonnet 4.6's 58.1% — and slightly outperforms Opus 4.8 on a knowledge work benchmark
- Anthropic is making Sonnet 5 the default model for free and Pro plans starting Tuesday
- The model shows a lower rate of "undesirable behaviors" than Sonnet 4.6, including better refusal of malicious requests and resistance to prompt-injection attacks, though it does not match Opus 4.8 or Claude Mythos Preview on misaligned-behavior evaluations
- Zapier engineer Daniel Shepard said the model completed a two-part Salesforce-plus-email job end-to-end that previously stalled, and Lovable co-founder Fabian Hedin praised its consistent refusals of unsafe requests
Why it matters: At $2 per million input tokens through August 31, Sonnet 5 undercuts Opus 4.8, GPT-5.5, and Gemini 3.1 Pro while approaching Opus-level performance on knowledge work — confirming, per the source, that agentic execution is now the baseline expectation at every tier, with cost and autonomous reliability as the remaining battleground. Anthropic is also pushing free-tier users onto the new model by default.
