---
title: "Claude 3.5 Sonnet Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Claude 3.5 Sonnet."
source: "https://rankedagi.com/models/claude-3.5-sonnet"
---

# Claude 3.5 Sonnet

| Field | Value |
| --- | --- |
| Organization | Anthropic |
| License | Proprietary |
| Version | Latest |
| Released | 2024-10-22 |
| Context window | 200K |
| Knowledge cutoff | 2024-04-01 |
| Input cost per million tokens | $3 |
| Output cost per million tokens | $15 |
| Cached input cost per million tokens | $0.3 |
| Model ID | claude-3-5-sonnet-20241022 |
| Last updated | 2026-05-04T01:34:22.840Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 41.6% | RankedAGI Coding Score |  |
| SWEBench Verified | coding | 49.0% | Agentic Coding | https://www.swebench.com/index.html |
| RankedAGI Agentic | Agents | 31.5% | RankedAGI Agentic Score |  |
| LiveCodeBench v6 | coding | 36.4% | LiveCodeBench v6 - Evaluation of Large Language Models for Code |  |
| LiveCodeBench v5 | coding | 39.8% | LiveCodeBench - Evaluation of Large Language Models for Code |  |
| ChatArena (LMSYS) | coding | 1313 | ChatArena (LMSYS) Coding ELO Score with Style Control |  |
| Codeforces ELO Score | coding | 717 | Codeforces ELO Score |  |
| Aider Polyglot | coding | 51.6% | Aider Polyglot Code Completion Benchmark |  |
| LiveBench Coding Score (old) | coding | 67.1% | LiveBench Coding Score old version |  |
| RankedAGI Reasoning | reasoning | 34.2% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 4.8% | Multidisciplinary Reasoning (no tools) |  |
| GPQA Diamond | reasoning | 65.0% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| Text Arena | general | 1355 | ChatArena (LMSYS) ELO Score | https://arena.ai/leaderboard/text |
| AIME 2025 | math | 3.0% | AIME 2025 Competition Math |  |
| AIME 2024 | math | 16.0% | AIME 2024 Competition Math |  |
| HumanEval | coding | 93.7% | HumanEval coding benchmark |  |
| HumanEval+ | coding | 86.2% | HumanEval+ benchmark |  |
| NYT Connections Extended | reasoning | 17.7% | NYT Connections Extended Version | https://github.com/lechmazur/nyt-connections/ |
| MMLU Pro | knowledge | 78.0% | Graduate-level visual problem-solving |  |
| MMLU | knowledge | 88.0% | Image Reasoning |  |
| MMMU | imaging | 70.4% | Multimodal Understanding College-level visual problem-solving |  |
| Hughes Hallucination Evaluation | safety | 4.6% | Computed using Vectara's Hallucination Evaluation Model, also known as HHEM. This evaluates how often an LLM introduces hallucinations when summarizing a document. | https://github.com/vectara/hallucination-leaderboard |
| AidanBench | general | 2691 | AidanBench Score |  |
| LiveBench Average | general | 60.7% | LiveBench Average Score old version |  |
| Instruction Following Evaluation | instruction-following | 89.3% | Instruction Following Evaluation |  |
| LiveBench Coding 25.4 | coding | 32.3% | LiveBench Coding Score 2025-04 | https://livebench.ai/ |
| LiveBench Data Analysis | general | 52.8% | LiveBench Data Analysis Score | https://livebench.ai/ |
| LiveBench Language | general | 53.8% | LiveBench Language Score | https://livebench.ai/ |
| Artificial Analysis Quality Index | general | 80 | Artificial Analysis Quality Index (third-party composite, deprecated) |  |
| RankedAGI Math | math | 36.3% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 40.4% | Overall RankedAGI score |  |

## Model Links

| Title | URL |
| --- | --- |
| Release | https://www.anthropic.com/claude/sonnet |