COMPARE.

anthropic

Claude 3 Haiku (20240307)

Arena evaluations

Model license

Proprietary

Arena 1194.7Human preference · 117,701 votes

anthropic

Claude 3 Opus (20240229)

Arena evaluations

Model license

Proprietary

Arena 1262.0Human preference · 194,909 votes

anthropic

Claude 3 Sonnet (20240229)

Arena evaluations

Model license

Proprietary

Arena 1218.0Human preference · 109,284 votes

anthropic

Claude 3.5 Haiku (20241022)

Arena evaluations

Model license

Proprietary

Arena 1255.4Human preference · 69,056 votes

anthropic

Claude 3.5 Sonnet (20240620)

Arena evaluations

Model license

Proprietary

Arena 1280.9Human preference · 82,419 votes

anthropic

Claude 3.5 Sonnet (20241022)

Arena evaluations

Model license

Proprietary

Arena 1297.8Human preference · 87,771 votes

anthropic

Claude 3.7 Sonnet (20250219)

Arena evaluations

Model license

Proprietary

Arena 1299.3Human preference · 42,573 votes

anthropic

Claude 3.7 Sonnet (Thinking 32K, 20250219)

Arena evaluations

Model license

Proprietary

Arena 1314.3Human preference · 38,156 votes

anthropic

Claude Haiku 4.5 (20251001)

Arena evaluations

Model license

Proprietary

Arena 1396.9Human preference · 154,063 votes

anthropic

Claude Opus 4 (20250514)

Arena evaluations

Model license

Proprietary

Arena 1366.0Human preference · 43,037 votes

anthropic

Claude Opus 4 (Thinking 16K, 20250514)

Arena evaluations

Model license

Proprietary

Arena 1377.1Human preference · 35,904 votes

anthropic

Claude Opus 4.1 (20250805)

Arena evaluations

Model license

Proprietary

Arena 1418.6Human preference · 77,213 votes

anthropic

Claude Opus 4.1 (Thinking 16K, 20250805)

Arena evaluations

Model license

Proprietary

Arena 1419.2Human preference · 49,535 votes

anthropic

Claude Opus 4.5 (20251101)

Arena evaluations

Model license

Proprietary

Arena 1450.9Human preference · 72,825 votes

anthropic

Claude Opus 4.5 (High, 32K, 20251101)

Arena evaluations

Model license

Proprietary

Arena 1448.2Human preference · 37,484 votes

anthropic

Claude Sonnet 4 (20250514)

Arena evaluations

Model license

Proprietary

Arena 1339.4Human preference · 39,168 votes

anthropic

Claude Sonnet 4 (Thinking 32K, 20250514)

Arena evaluations

Model license

Proprietary

Arena 1350.9Human preference · 34,128 votes

anthropic

Claude Sonnet 4.5 (20250929)

Arena evaluations

Model license

Proprietary

Arena 1438.5Human preference · 82,621 votes

anthropic

Claude Sonnet 4.5 (High, 32K, 20250929)

Arena evaluations

Model license

Proprietary

Arena 1434.5Human preference · 83,892 votes

meta

Code Llama 34B Instruct

Arena evaluations

Model license

Llama 2 Community

Arena 1065.5Human preference · 7,366 votes

meta

Code Llama 70B Instruct

Arena evaluations

Model license

Llama 2 Community

Arena 1057.2Human preference · 1,143 votes

cohere

Command A (03-2025)

Arena evaluations

Model license

CC-BY-NC-4.0

Arena 1330.9Human preference · 55,763 votes

cohere

Command R (08-2024)

Arena evaluations

Model license

CC-BY-NC-4.0

Arena 1187.3Human preference · 10,140 votes

cohere

Command R

Arena evaluations

Model license

CC-BY-NC-4.0

Arena 1163.1Human preference · 54,036 votes