Overview
Claude Sonnet 5 is a mid-range flagship model quietly released by Anthropic on **June 30, 2026**—without a grand launch event or prior announcement, directly available on Claude Code and Claude Platform. The official positioning is highly ambitious: **"Near-Opus thinking, at Sonnet speed"** (close to Opus 4.8's reasoning ability while retaining Sonnet's speed and price). This is why Anthropic calls it "**the most Agent-capable Sonnet model to date**"—Anthropic's strategy is to compress the expensive Opus-level Agent capabilities into the mid-range Sonnet tier, using price advantages to reclaim market share eroded by competitors such as GPT-5.6, Gemini, Kimi K3, and DeepSeek V4.
**Key Pricing (the biggest selling point)**:
- **Limited-time price** (until August 31, 2026): Input **$2/M tokens**, Output **$10/M tokens**—compared to Sonnet 4.8's $3/$15, a reduction of about 33-40%
- **Regular price** (from September 1): Will increase, but still significantly lower than Opus 4.8
- **Cursor official data**: CursorBench 3.1 score **57%** (Sonnet 4.6 was 49%), a clear capability upgrade
**⚠️ Launch Day Controversy: Agent Capabilities Surpassed by Domestic Models**: Within 24 hours of Sonnet 5's release, there was a **flood of negative reviews**. Community tests revealed:
- In some Agent tasks, it **"cannot beat Qwen 3.7-Max and MiniMax M3"**, especially in long-chain reasoning and tool call stability
- Some developers reported that Sonnet 5 is "less stable than Sonnet 4.6" when handling complex projects, creating a gap between advertised specs and actual experience
- Official benchmarks like SWE-bench have not yet released full data, and the community is calling for more transparent evaluations from Anthropic
This wave of negative feedback has been interpreted by the industry as "**Anthropic, thrown off balance after GPT-5.6's release, rushed Sonnet 5's launch without sufficient polish**." However, some developers' tests suggest that Sonnet 5 is indeed stronger than 4.8 in **daily primary coding tasks**, but expectations were inflated by the official "near Opus 4.8" marketing.
**Latest Update (July 24, 2026)**: Anthropic released **Opus 5**, positioned as a major upgrade in the Opus tier, emphasizing stronger coding abilities, smarter Agent capabilities, and sharper professional performance. The launch of Opus 5 further enriches Anthropic's product line, forming a high-low combination with Sonnet 5. The limited-time price deadline for Sonnet 5 (August 31, 2026) remains valid, but users can now also choose Opus 5 for more complex tasks.
**Current Positioning**: If you prioritize cost-effectiveness and stable daily use, Sonnet 5 is worth trying; if your scenario involves complex long-chain Agent tasks, you may need to conduct actual tests comparing Sonnet 5, Opus 5, GPT-5.6 Sol, Kimi K3, and Tongyi Qwen 3.7-Max before deciding. This also confirms the industry theme for the second half of 2026: **the cost-effectiveness battle for mid-range Agent models, where Anthropic is no longer the sole leader**.
Key Features
- "Near-Opus thinking, at Sonnet speed" positioning: Official core slogan: reasoning ability close to Opus 4.8, retaining Sonnet's speed and price
- Limited-time price significantly reduced (until Aug 31): $2/M input + $10/M output, about 33-40% reduction compared to Sonnet 4.8
- Significantly improved Agent capabilities: Officially called "the most Agent-capable Sonnet to date"; Cursor's CursorBench 3.1 score is 57% (4.6 was 49%)
- Three major upgrades: coding + tool use + reasoning: Comprehensive improvements in tool call accuracy, multi-step reasoning stability, and code generation quality
- Already available on Claude Code and Platform: Developers can directly call via model ID `claude-sonnet-5`; major AI coding tools like Cursor and Cline have quickly integrated it
- ⚠️ Launch day controversy: Community feedback indicates some Agent tasks are surpassed by Qwen 3.7-Max and MiniMax M3; concerns about long-chain reasoning and tool call stability
- Opus 5 released (July 24, 2026): Anthropic released Opus 5, positioned as a major upgrade in the Opus tier, emphasizing stronger coding, smarter Agent capabilities, and sharper professional performance, forming a high-low combination with Sonnet 5
Pros
- Reasoning ability close to Opus 4.8, while maintaining Sonnet's speed and price
- Limited-time price significantly reduced ($2/$10 per M token), outstanding cost performance
- Agent capabilities significantly improved, CursorBench 3.1 score 57%
- Comprehensive upgrades in programming, tool calling, and reasoning, suitable as a daily development主力
- Already available on Claude Code and mainstream tools, easy integration
Summary
Claude Sonnet 5 is a precisely positioned mid-range flagship model, offering reasoning capabilities close to Opus 4.8 while maintaining Sonnet-level speed and pricing. It focuses on cost-effective daily programming and Agent tasks. A limited-time price reduction of 33-40% makes it one of the most attractive mid-range options in the second half of 2026, especially for developers seeking a stable primary model. However, it should be noted that it may be outperformed by domestic models in complex long-term Agent tasks, so comparative testing based on actual scenarios is recommended.
Version History
- 官网于2026年6月30日发布全新模型Claude Sonnet 5，定位为最具代理能力（agentic）的Sonnet: The official website released the new model Claude Sonnet 5 on June 30, 2026, positioning it as the most agentic model in the Sonnet series, with performance close to Opus 4.8 but at a lower price. Current evaluations do not mention this version.
- 官网发布了全新模型Claude Sonnet 5（2026年6月30日），定位为最具代理能力（agentic）的Sonn: The official website has released a new model, Claude Sonnet 5 (June 30, 2026), positioned as the most agentic Sonnet model, with performance close to Opus 4.8 but at a lower price, supporting planning, tool use (browser/terminal), and autonomous operation. Current evaluations do not mention this version.