Claude Sonnet 4.6: Opus-Level Coding Performance at Sonnet Pricing
Anthropic releases Claude Sonnet 4.6 on February 17, achieving near-Opus benchmark scores at five times lower cost, with developers preferring it over previous Opus models.
First Sonnet to Beat Previous Opus
Anthropic released Claude Sonnet 4.6 on February 17, 2026, at the same price as Sonnet 4.5 — but with performance that rivals the much more expensive Opus tier. For the first time ever, a Sonnet model is preferred over the previous generation's Opus in coding evaluations.
Benchmark Results
| Benchmark | Sonnet 4.6 | Sonnet 4.5 | Opus 4.6 |
|---|---|---|---|
| SWE-bench Verified | 79.6% | 77.2% | ~80% |
| OSWorld (Computer Use) | 72.5% | — | 72.7% |
| ARC-AGI-2 | 58.3% | 13.6% | — |
| Terminal-Bench 2.0 | 59.1% | — | — |
The ARC-AGI-2 score represents a massive 4.3x improvement over Sonnet 4.5, jumping from 13.6% to 58.3%.
Developer Preference
In Claude Code testing, developers preferred Sonnet 4.6 over Sonnet 4.5 70% of the time and over the previous flagship Opus 4.5 59% of the time. This is unprecedented for a Sonnet-class model.
Pricing and Context
At $3/$15 per million tokens — five times cheaper than Opus — Sonnet 4.6 also introduces a 1 million token context window (beta), making it the first Sonnet-class model to support full codebase analysis in a single prompt.
The model is available across Claude.ai, the API, Amazon Bedrock, and Microsoft Foundry.
Source: CNBC | The New Stack | DataCamp | SitePoint
NYC News
Expert insights on AI voice agents and customer communication automation.
Try CallSphere AI Voice Agents
See how AI voice agents work for your industry. Live demo available -- no signup required.