Recent studyBlind TTS Elo is live. Compare two anonymous voice samples, vote after listening, and help separate real preference signal from noise.Vote in the study ->
Codesota · Models · Llama 3 (405B, Instruct)Meta11 results · 10 benchmarks
Model card

Llama 3 (405B, Instruct).

MetaLarge language model1 current SOTA

Imported from Papers with Code dump

§ 02 · Benchmarks

Every benchmark Llama 3 (405B, Instruct) has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01HumanEvalComputer Code · Code Generationpass-189.0%#1/3source ↗
02HumanEval+Computer Code · Code Generationpass-189.0%#1/12source ↗
03MBPPComputer Code · Code Generationpass-178.8%#3/3source ↗
04MBPP+Computer Code · Code Generationpass-178.8%#4/9source ↗
05GSM8KReasoning · Mathematical Reasoningaccuracy96.8%#18/48source ↗
06LibriSpeechSpeech · Speech Recognitionwer3.1%#28/92source ↗
07MATHReasoning · Mathematical Reasoningaccuracy73.8%#31/46source ↗
08MMLUReasoning · Commonsense Reasoningaccuracy87.3%#31/64source ↗
09MMLU-ProReasoning · Commonsense Reasoningaccuracy73.3%#55/732024-07-31source ↗
10MMLU-ProReasoning · Commonsense Reasoningaccuracy73.3%#55/73source ↗
11GPQA DiamondReasoning · Multi-step Reasoningaccuracy51.1%#61/74source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area

Where Llama 3 (405B, Instruct) actually performs.

Computer Code
4
benchmarks
avg rank #2.3 · 1 SOTA
Speech
1
benchmark
avg rank #28.0
Reasoning
5
benchmarks
avg rank #41.8
§ 04 · Papers

1 paper with results for Llama 3 (405B, Instruct).

  1. 2024-07-31· 10 results

    The Llama 3 Herd of Models

§ 05 · Related models

Other Meta models scored on Codesota.

Llama 3 70B
9 results
Llama 3.1 405B
7 results
Llama-4-Maverick
400B total / 17B active (128 experts) params · 6 results
Llama 3.1 70B
4 results
Code Llama 34B
Unknown params · 2 results
ConvNeXt V2 Huge
650M params · 2 results
DeiT-B Distilled
86M params · 2 results
CodeLlama 70B
70B params · 1 result
§ 06 · Sources & freshness

Where these numbers come from.

pwc-dump
10
results
paperswithcode
1
result
0 of 11 rows marked verified.