Claude Sonnet 3.5 v2

Anthropic · anthropic/claude-3-5-sonnet-20241022

ProviderAnthropic
Familyclaude-sonnet
Typevlm
Statusactive
Released2024-10-22
Updated2024-10-22

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
AI2D (AI2 Diagrams)unassessed80.2
Aider Polyglot Benchmarkunassessed51.6
Arena Elo — Codingunassessed1310.0
Arena Elo — Hard Promptsunassessed1396.7
Arena Elo — Mathunassessed1290.0
Arena Elo — Overall (Text)unassessed1300.0
Arena Elo — Style Controlunassessed1371.4
Arena Elo — Visionunassessed1160.4
BBQ (Bias Benchmark for QA)unassessed87.2
ChartQAunassessed90.8
DocVQAunassessed95.2
FLORES English-to-Germanunassessed66.9
FLORES English-to-Spanishunassessed70.8
FLORES English-to-Japaneseunassessed60.8
FLORES English-to-Chineseunassessed59.2
GPQA Diamondunassessed65.0
GSM8Kunassessed96.4
HELM Safetyunassessed91.5
HumanEvalunassessed86.8
IFEvalunassessed85.4
LiveCodeBenchunassessed38.1
MATH-500unassessed78.3
MathVistaunassessed67.7
MGSM (Multilingual Grade School Math)unassessed91.6
MMLU-Prounassessed76.2
MMMUunassessed65.9
MultiPL-E: C#unassessed80.2
MultiPL-E: Juliaunassessed61.5
MultiPL-E: Kotlin (unconfirmed)unassessed72.8
MultiPL-E: Luaunassessed55.8
MultiPL-E: Perlunassessed52.2
MultiPL-E: PHPunassessed77.5
MultiPL-E: Runassessed58.2
MultiPL-E: Rubyunassessed68.5
MultiPL-E: Scalaunassessed62.5
MultiPL-E: Swiftunassessed66.5
OCRBenchunassessed78.8
SWE-bench Verifiedunassessed49.0
ToxiGenunassessed94.5

Data

This card as JSON · Edit on GitHub