On June 30, 2026, Anthropic officially released its new model, Claude Sonnet 5. The company calls it its "most agentic Sonnet yet" and says it delivers performance close to the higher-end Opus 4.8 at a lower price.
June 30, 2026 · Anthropic
Claude Sonnet 5: Opus-level performance at Sonnet pricing
Anthropic's "most agentic Sonnet model" brings planning, tool use and autonomous task execution that approaches Opus-class levels — now the default on Free and Pro, with a 1M-token context window.
Default
model on Free & Pro plans
−60%
cheaper output vs Opus 4.8
API pricing — output, per million tokens
The cost story: near-Opus capability, well below Opus rates
Input rates follow the same gap: $2 / $3 / $5 per million tokens (intro / standard / Opus).
What partners praised
Sustained coding & debugging across complex multi-step software work
Completed end-to-end a two-part job that previously stalled midway
Writes, fixes & verifies reproducing tests in a single pass, unprompted
Strong root-cause tracing in brownfield code
Caveats to watch
A gap with Opus 4.8 remains at high effort levels
Opus retains an edge on certain benchmarks
Lags significantly on cybersecurity tasks — unsafe cyber use blocked by default
Tokenizer update may shift token counts; real-world cost varies by usage
Availability today
Live now in Claude Code and the Claude API (model ID claude-sonnet-5 ), rolling out to AWS Bedrock , Azure and Google Vertex . Introductory pricing runs through Aug 31, 2026.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…