Orbit/Models/Claude Sonnet 4.5

Claude Sonnet 4.5

Context
200K tokens
Input
$3/M
Output
$15/M
Arena
1332
License
Proprietary
Released
Sep 2025

Overview

Claude Sonnet 4.5 is a proprietary AI model developed by Anthropic, released on September 29, 2025. It is part of the Claude 4.5 family and is designed to handle both text and image inputs, offering a context window of 200,000 tokens. This model is fine-tuned for general tasks and is particularly optimized for coding applications, demonstrating state-of-the-art performance on various coding benchmarks. Notably, it has achieved an Arena Elo score of 1,332, indicating strong performance in competitive settings. Intended for tasks requiring extensive context and multimodal input, Claude Sonnet 4.5 is well-suited for applications in coding, real-world agents, and computer use. Its strengths include a large context window, multimodal capabilities, and competitive pricing, making it a cost-effective choice for developers. However, it may lag behind newer models like Claude Sonnet 4.6 in certain areas such as coding and tool use. In comparison to peers, Claude Sonnet 4.5 offers a balance between performance and cost, positioning itself as a reliable option for various AI applications. The model is available through the Claude API and the Claude chatbot, with pricing set at $3 per million input tokens and $15 per million output tokens. This pricing structure aligns with that of its predecessor, Claude Sonnet 4, and is consistent across different deployment routes. As of June 2026, Claude Sonnet 4.5 remains a competitive choice for developers seeking a robust AI model for coding and related tasks.

Strengths

codinglong contextreasoningmath

Benchmarks

Arena Elocoding1332
HumanEvalcoding86.1 / 100
MMLU PROreasoning86 / 100
GPQA Diamondreasoning56.2 / 100