Claude Sonnet 4.5

Anthropic · anthropic/claude-sonnet-4-5-20250929

ProviderAnthropic
Familyclaude-sonnet
Typellm-reasoning
Statusactive
Released2025-09-29
Updated2025-09-29

Reported benchmark scores

These scores carry one collection date for the whole card (2026-04) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.
BenchmarkCatalogue standingScore
Aider Polyglot Benchmarkunassessed75.5
AlpacaEvalunassessed50.8
Arena Elo — Codingunassessed1380.0
Arena Elo — Hard Promptsunassessed1485.0
Arena Elo — Mathunassessed1360.0
Arena Elo — Overall (Text)unassessed1370.0
Arena Elo — Style Controlunassessed1451.9
Artificial Analysis Intelligence Indexunassessed86.0
Artificial Analysis Output Speedunassessed71.0
BBQ (Bias Benchmark for QA)unassessed87.5
ChartQAunassessed84.2
DocVQAunassessed91.5
GPQA Diamondunassessed72.5
HELM Safetyunassessed91.8
HumanEvalunassessed90.8
IFEvalunassessed89.3
LiveCodeBenchunassessed59.0
MATH-500unassessed92.1
MathVistaunassessed62.1
MMLU: Astronomyunassessed77.8
MMLU: Biology (subcategory)unassessed85.5
MMLU: Business Ethicsunassessed79.5
MMLU: Chemistry (subcategory)unassessed79.8
MMLU: Clinical Knowledgeunassessed84.2
MMLU: Computer Science (subcategory)unassessed86.2
MMLU: Jurisprudenceunassessed77.8
MMLU: Physics (unresolved key)unassessed82.5
MMLU-Prounassessed81.7
MMLU: Professional Accountingunassessed69.8
MMLU: Professional Lawunassessed75.5
MMMUunassessed68.2
MT-Benchunassessed9.2
MultiPL-Eunassessed86.7
MultiPL-E: C++unassessed85.8
MultiPL-E: C#unassessed85.5
MultiPL-E: Gounassessed82.8
MultiPL-E: Javaunassessed89.1
MultiPL-E: JavaScriptunassessed89.5
MultiPL-E: Juliaunassessed67.5
MultiPL-E: Kotlin (unconfirmed)unassessed79.8
MultiPL-E: Luaunassessed62.1
MultiPL-E: Perlunassessed58.8
MultiPL-E: PHPunassessed82.2
MultiPL-E: Pythonunassessed93.2
MultiPL-E: Runassessed65.2
MultiPL-E: Rubyunassessed75.2
MultiPL-E: Rustunassessed78.5
MultiPL-E: Scalaunassessed68.5
MultiPL-E: Swiftunassessed73.5
MultiPL-E: TypeScriptunassessed88.2
SWE-bench Agentunassessed55.8
SWE-bench Verifiedunassessed62.1
τ-benchunassessed61.5
Terminal-Benchunassessed48.2
ToxiGenunassessed94.5
WildBenchunassessed78.2

Data

This card as JSON · Edit on GitHub