Breaking

Anthropic Launches Claude Sonnet 5, Bringing Near-Opus Ability at Lower Prices

July 1, 2026 at 06:42 EDT

Foundation Models
AI Agents
Software Dev & Coding

On June 30, 2026, Anthropic officially announced and released a new model, "Claude Sonnet 5" (model ID: claude-sonnet-5), saying it delivers agentic performance—previously requiring the larger, pricier Opus class only months ago—at Sonnet-tier pricing.

June 30, 2026 · Anthropic

Claude Sonnet 5: Opus-Class Power in the Mid-Tier Band

Anthropic's "most agentic Sonnet yet" ships as a drop-in successor to Sonnet 4.6 — handling planning, tool use and autonomous execution at a level once reserved for larger, pricier models.

token context (128k output)

$2 / $10

launch input/output per 1M tokens

1.0–1.35×

token counts vs old tokenizer

Launch pricing vs standard rate

Per million tokens — promo runs through Aug 31, 2026, then reverts to Sonnet 4.6 levels.

INPUT

Launch

Standard

OUTPUT

$10

Launch

$15

Standard

What changed under the hood

Adaptive thinking on by default

Manual extended thinking now returns a 400 error.

New tokenizer

Prompt caching & Batch API discounts still apply.

Broad day-one reach

Claude.ai, API, Bedrock, Vertex AI, GitHub Copilot.

Developers praise

Higher completion rates on agentic tasks
Self-verifying behavior on CLI work
1M context fits long, large-scale jobs

Critics push back

Opus 4.8 seen as cheaper and better by some
More token usage and hallucinations reported
No Priority Tier support

Continue reading

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