Breaking

Anthropic Launches Claude Sonnet 5, Nearing Opus 4.8 at Lower Cost

June 30, 2026 at 14:41 EDT

Foundation Models
AI Agents
Software Dev & Coding

Anthropic officially unveiled its new model "Claude Sonnet 5" on June 30, 2026, calling it "the most agentic Sonnet model yet" and offering performance close to its higher-end Claude Opus 4.8 at a lower price. At launch it became the default model on the Free and Pro plans, and is immediately available to Max/Team/Enterprise users, via Claude Code, the Claude API (model ID claude-sonnet-5), and AWS Bedrock.

June 30, 2026 · Anthropic

Claude Sonnet 5: Opus-Class Agentic Power at Sonnet Prices

Anthropic's new balanced flagship lands across all Claude.ai plans, the API, and AWS — pushing frontier coding and agentic capability into a lower price tier, now the default model on Free and Pro.

63.2%

SWE-Bench Pro
(up from 58.1%)

token context
window (default)

128k

max output
tokens

$2/$10

intro input/output
per M tokens

SWE-Bench Pro — harder engineering tasks

Practical coding & agentic execution lifts from Sonnet 4.6 to Sonnet 5.

58.1%

Sonnet 4.6

63.2%

Sonnet 5

What changed under the hood

Adaptive thinking by default

Manual extended thinking removed — specifying it returns a 400 error.

New tokenizer

Same text now uses ~30% more tokens — factor into cost.

Fixed sampling

Non-default temperature/top_p/top_k return a 400 error.

Limited cybersecurity

Capabilities intentionally restricted. No Priority Tier support.

Developers say

Agentic performance near Opus 4.8 at Sonnet pricing.
Less stalling in planning, browser, terminal and long-running tasks.
Cost-efficient enough to be a main daily-workflow model.

Caveats raised

Opus 4.8 still leads on certain tasks.
Limiting cybersecurity could make vulnerable code more likely.
New tokenizer raises effective token spend ~30%.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.