Model

Claude 3.5 Sonnet

CLOSED

Anthropic · Claude 3.5 family · released Jun 20, 2024

Mid-tier that beat Opus 3, launched the Artifacts feature, and led coding/agentic benchmarks at a fraction of Opus's price.

ReasoningCodingVisionFunction callingTool useAgentic

1294.7

Elo · rank #156

Parameters

Undisclosed

Active params

—

Context

200K tokens

Architecture

Dense multimodal Transformer (parameter count undisclosed)

License

Proprietary (Anthropic Commercial Terms of Service)

Languages

—

API price (in/out)

$3 / $15

Modalities

text · vision

Benchmark results

Bar shows position within the tracked field; marker = field best

AI2DVision94.7%#7

best: Molmo 72B · 96.3%

AIMEMath9.6%#147

best: GPT-5.2 · 100.0%

Arena EloHuman preference1342#31

best: Claude Fable 5 · 1505

BIG-Bench HardReasoning93.1%#3

best: ERNIE 4.5 300B-A47B · 94.3%

ChartQAVision90.8%#2

best: MiniMax-VL-01 · 91.7%

DocVQAVision95.2%#7

best: Qwen2-VL-72B · 96.5%

DROPReasoning87.1#12

best: Hunyuan-T1 · 93.1

GPQA DiamondReasoning59.4%#167

best: GPT-5.6 · 94.6%

GSM8KMath96.4%#4

best: Llama 3.1 405B · 96.8%

HumanEvalCoding92.0%#14

best: Claude Opus 4.5 · 99.4%

IFEvalReasoning87.8%#35

best: Gemma 4 26B A4B · 98.5%

LogicKorHuman preference9.21#3

best: GPT-4o · 9.33

MATH-500Math71.1%#87

best: GPT-5 · 99.4%

MathVistaVision67.7%#33

best: Seed 2.1 Pro · 90.7%

MGSMMath91.6%#5

best: OpenAI o4-mini · 93.7%

MMLU-ProKnowledge75.1%#76

best: Claude Fable 5 · 91.5%

MMLUKnowledge88.7%#18

best: OpenAI o3 · 92.9%

MMMUVision68.3%#61

best: Claude Fable 5 · 89.3%

SimpleQAKnowledge35.0%#12

best: GPT-4.5 · 62.5%

StreamingBenchVision59.7%#4

best: Gemini 1.5 Pro · 70.3%

SWE-bench VerifiedCoding33.0%#109

best: Claude Fable 5 · 95.0%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

Output / M tok

$15

Claude 3.5 family

Elo progression across releases

Claude 3.5 SonnetJun 20241294.7 Claude 3.5 HaikuOct 20241041.8 Claude 3.5 Sonnet (New)Oct 20241357.3

Resources

DOCSdocs.claude.com

API price $3/$15 · each benchmark row carries its own source badge (see methodology)