Claude Sonnet 5
Claude Sonnet 5 is Anthropic's Sonnet-family model released on June 30, 2026, with the API model ID claude-sonnet-5 and a focus on coding, agents, and long-context reasoning.
Sonnet is Anthropic's practical workhorse tier, so a Sonnet 5 release affects everyday model routing more than a niche flagship update. Teams need to compare its coding quality, 1M-token context, 128K max output, tool-use behavior, introductory $2/$10 pricing through August 31, 2026, and standard $3/$15 pricing against Opus, Fable, GPT, and lower-cost model routes.
Anthropic announced Claude Sonnet 5 on June 30, 2026 and describes it as a coding- and agent-oriented upgrade over earlier Sonnet models. The current official model overview lists the dateless pinned model ID, context window, max output, modalities, platform availability, and pricing. The Sonnet 5 release notes and migration guide add behavior changes such as high effort defaults, improved tool use, unsupported temperature control, and prompt-migration caveats. Hacker News discussion adds reader demand around price, coding value, access, and whether Sonnet 5 should replace Opus or other frontier routes.
- Default coding-agent model trials for Claude Code and Anthropic API workflows.
- Long-context software engineering tasks that need 1M-token project context.
- Agent workflows with tool use, bash execution, code editing, and parallel tool patterns.
- Cost-aware routing between Sonnet, Opus, Fable, GPT, GLM, and DeepSeek model choices.
- Migration testing from Sonnet 4.5, Sonnet 4.6, or older Claude 4.x deployments.
Anthropic positions Claude Sonnet 5 as a Sonnet-family model for coding, agentic work, and production use. Current official docs list claude-sonnet-5 as a dateless pinned model ID, a 1M-token context window, 128K maximum output, text and image input, text output, and availability through Anthropic API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.
- Release date: June 30, 2026.
- Model ID: claude-sonnet-5.
- Introductory pricing: $2 per million input tokens and $10 per million output tokens through August 31, 2026.
- Standard pricing from September 1, 2026: $3 per million input tokens and $15 per million output tokens.
Treat Sonnet 5 as the likely default candidate for hard everyday coding work, then escalate or downshift only after measuring your own harness. Anthropic highlights stronger coding and tool-use behavior, but the migration guide also warns that some prompt behavior changes need validation before a production switch.
- Compare against Sonnet 4.5 or Sonnet 4.6 for routine coding cost and latency.
- Compare against Opus 5 or Fable 5 when long-horizon task quality matters more than token cost.
- Re-test prompts that rely on temperature, deep thinking controls, tool formats, or cached behavior.
Early discussion focuses less on generic chatbot quality and more on operational routing: whether the higher Sonnet price is justified, whether the coding jump is visible inside Claude Code or API harnesses, and when teams should keep Opus/Fable-class models as escalation options.
Anthropic Opus-tier model used as a high-capability escalation comparison.
Anthropic public high-capability model for stronger long-horizon workflows.
OpenAI generally available model family for frontier, balanced, and cost-sensitive coding and reasoning work.
Long-context coding model used as an open-model and provider-cost comparison point.
Source confidence
Claude Sonnet 5 FAQ
Common questions about Claude Sonnet 5.
What is the Claude Sonnet 5 model ID?+
The official Anthropic API model ID is claude-sonnet-5. Anthropic documents this dateless ID as a pinned snapshot rather than an evergreen alias.
Is Claude Sonnet 5 better for coding than Sonnet 4.5?+
Anthropic positions Sonnet 5 as a major coding and agent upgrade over earlier Sonnet releases, but teams should still validate it on their own repositories, tools, tests, and cost targets before replacing Sonnet 4.5 or Sonnet 4.6.
How much does Claude Sonnet 5 cost?+
Anthropic lists introductory Claude Sonnet 5 pricing at $2 per million input tokens and $10 per million output tokens through August 31, 2026. Standard pricing becomes $3 per million input tokens and $15 per million output tokens on September 1, with separate prompt-caching and batch-processing rates.
Should Sonnet 5 replace Opus or Fable models?+
Not automatically. Sonnet 5 is a strong default candidate for many coding and agent tasks, while Opus or Fable-style models may still be worth the higher cost for the hardest long-horizon reasoning, review, or escalation workflows.