# Claude Sonnet 4.5 Review & Benchmarks (2026): Price, Context, Specs | Alher Tech

> Claude Sonnet 4.5 by Anthropic: benchmarks, pricing, context window, strengths and use cases. Data from official benchmark sources.

- Canonical page: https://alhertech.com/en/ai-comparison/claude-sonnet-4-5/
- Site: Alher Tech (custom software, AI agents and SEO engineering, https://alhertech.com/)
- Contact: https://alhertech.com/en/contact/

---

Last updated: April 22, 2026

## Specs

- **Vendor**: Anthropic
- **Released**: June 12, 2025
- **Context window**: 200K tokens
- **Input**: $3 / M tokens
- **Output**: $15 / M tokens

## Benchmarks

- **mmlu**: 86.9
- **gpqa**: 56.2
- **humanEval**: 86.1
- **arenaElo**: 1332
- **sweBench**: 58
- **sweBenchPro**: 50
- **liveCodeBench**: 68
- **aiderPolyglot**: 70
- **terminalBench**: 40
- **mmluPro**: 78
- **hle**: 14
- **aime**: 76
- **math500**: 92
- **mmmu**: 72

## Strengths

- Cheaper than Opus 4.6 with very close quality
- Still the most integrated Sonnet in third-party tooling
- Predictable latency on Anthropic infra

## Weaknesses

- Behind Sonnet 4.6 on coding and tool use
- Same 200K context: no gain on long-doc workloads

## Best for

- Projects already in production on Sonnet 4.5
- Customers waiting to validate Sonnet 4.6 before upgrade
- Benchmark baselines for migration decisions

[Compare with other AI models](https://alhertech.com/en/ai-comparison/)