via TechCrunch AI
Anthropic Launches Claude Sonnet 5: A Cheaper Path to Autonomous Agents
As agentic capabilities become the baseline expectation among foundation model companies, Anthropic has released Claude Sonnet 5 — a more powerful and agentic version of its midsize model. The move signals that the competitive battleground is shifting from who can deliver agentic features to who can offer them most affordably and reliably.
“Sonnet 5 can make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models,” Anthropic said in a blog post.
This framing mirrors the messaging from OpenAI and Google on their recent releases. OpenAI’s GPT-5.6 Sol, launched in preview last week, is also the company’s most agentic model yet, enabling users to split work across subagents for longer autonomous tasks. Google’s Gemini 3.5 Flash, which debuted in May, was pitched as a pivot from conversational chatbot to agentic tool that plans, builds, and iterates on real work with minimal human input.
Sonnet 5’s pitch confirms that agentic capability is now the default expectation at every price tier. The key differentiator is no longer who can deliver the best agentic performance, but who can do it at the lowest cost and with the least need for human oversight.
Sonnet 5 promises performance close to that of Opus 4.8 — Anthropic’s most capable model — but at a fraction of the price. Starting Tuesday, Claude Sonnet 5 becomes the default model for free and Pro plans and is available across all subscription tiers.
At launch, Sonnet 5 is priced at $2 per million input tokens and $10 per million output tokens through August 31, after which the price rises to $3 per million input tokens and $10 per million output tokens. This makes Sonnet 5 cheaper than Opus 4.8, OpenAI’s GPT-5.5, and Google’s Gemini 3.1 Pro. It remains slightly more expensive than Gemini 3.5 Flash, however.
According to Anthropic, Sonnet 5 shows significant improvements over its predecessor, Sonnet 4.6 (released in February 2026), across key agentic dimensions such as reasoning, tool use, software coding, and knowledge work.
Benchmark results illustrate the gains: Sonnet 5 scores 63.2% on agentic coding, compared to Opus 4.8’s 69.2% and Sonnet 4.6’s 58.1%. On a knowledge work benchmark, Sonnet 5 actually edges past Opus 4.8 — a model known for excellence on hard problems like nuanced judgment calls and deep research.
“Opus 4.8 is still the model of choice for higher accuracy on these tasks, but Sonnet 5 provides developers with lower-priced options that are of much higher quality than what was previously available,” Anthropic notes. “Between Sonnet 5 and Opus 4.8, users can adjust effort level to find the right balance of cost and performance.”
Early testers cited by Anthropic report that Sonnet 5 excels at completing complex tasks where earlier models would have halted prematurely. The model also tends to “check its own output without being explicitly asked,” improving reliability in autonomous workflows.
With the 2026 AI landscape increasingly defined by agentic workflows, Claude Sonnet 5 positions Anthropic to compete not just on raw capability, but on cost-effective autonomy — a move that may pressure rivals to further cut prices or risk losing developer adoption.