Développé par Anthropic, Claude Sonnet 4.5 est un modèle de langage de pointe qui excelle dans le raisonnement en plusieurs étapes, la programmation avancée, l'utilisation de l'ordinateur (Computer Use) et des domaines spécialisés tels que la finance.

Pricing: Input 170 credits / 1M tokens (≈ $0.850, ~28.3% of official pricing), Output 855 credits / 1M tokens (≈ $4.275, ~28.5% of official pricing). High-tier top-ups include +10% bonus credits. Effective pricing is approximately 10% lower than the above rates.
claude-sonnet-4-5

Automatically append previous messages to maintain multi-turn context. May increase token usage.

Structured outputs
Function calling
Web Access
Thinking
Reasoning effort
Max tokens
Stream Response

How can I help you today?

README

Complete guide to using claude-sonnet-4-5

API Claude Sonnet 4.5 à prix abordable

Image de démonstration de la bannière principale montrant les composants de l'interface
Performances en ingénierie logicielle et codage agentique
Computer Use et benchmarks de raisonnement multi-domaines
Fonctionnalité / BenchmarkClaude Sonnet 4.5 (Anthropic)Gemini 3 Pro (Google)
Primary StrengthSoftware Engineering & Complex AgentsMultimodal Reasoning & Massive Context
Input Context Window200K (up to 1M in beta)1M (standard)
Max Output TokensUp to 64K64K
Multimodal SupportText, ImageText, Image, Audio, Video, PDF
Coding Accuracy (SWE-bench Verified)77.2%76.2%
Computer-Use Tasks (OSWorld)61.4%~54.2%
Math Reasoning (AIME)82.5%84.0%
General Knowledge (MMLU)89.1%91.9%
Ideal Use CaseIDE extensions, autonomous coding agents, GUI automationVideo analysis, large-scale document research, Google ecosystem apps
4.8/ 5
51,212 ratingsSlide your rating
PoorExcellent