GeekRadar / The comparison desk

COMPARE.

AI models

Code Llama 70B Instruct

Model evaluation

Capability & preference

Human preference · Arena

Overall text preferences from head-to-head votes. Not factual accuracy or a task-success rate.

Code Llama 70B Instruct 1057.2 Code Llama 70B Instruct · 95% interval 1039.2–1075.2 · 1,143 votes · 2026-10-28

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

Arena text preference / englishhuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,095.87 rating

Confidence interval: 1,074.02 – 1,117.72

695 votes

Oct 28, 2026

Category: english
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

872.24 rating

Confidence interval: 844.91 – 899.57

809 votes

Oct 28, 2026

Category: exclude_ties
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / hard_promptshuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,052.02 rating

Confidence interval: 1,021.54 – 1,082.51

306 votes

Oct 28, 2026

Category: hard_prompts
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,077.49 rating

Confidence interval: 1,041.38 – 1,113.61

207 votes

Oct 28, 2026

Category: hard_prompts_english
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / industry_software_and_it_serviceshuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,057.50 rating

Confidence interval: 1,026.00 – 1,089.01

325 votes

Oct 28, 2026

Category: industry_software_and_it_services
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / industry_writing_and_literature_and_languagehuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,014.71 rating

Confidence interval: 978.24 – 1,051.18

262 votes

Oct 28, 2026

Category: industry_writing_and_literature_and_language
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / instruction_followinghuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,023.54 rating

Confidence interval: 993.97 – 1,053.11

358 votes

Oct 28, 2026

Category: instruction_following
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / non_englishhuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

992.19 rating

Confidence interval: 964.35 – 1,020.02

448 votes

Oct 28, 2026

Category: non_english
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Arena text preference / overallhuman preference · 2026-10-28 · rating

Code Llama 70B Instruct

1,057.22 rating

Confidence interval: 1,039.24 – 1,075.19

1143 votes

Oct 28, 2026

Category: overall
Evaluated configuration: Code Llama 70B Instruct
Evaluation type: human preference

Run it locally

Inference · memory estimate

Code Llama 70B Instruct

Not enough verified architecture information for a reliable VRAM estimate.

Capabilities

Model license: Llama 2 Community