Llama 3.1 405B — Benchmarks

Benchmark scores for Llama 3.1 405B aggregated from public leaderboards (8 boards), with how it ranks among open models and how it compares with its nearest neighbours. See hardware requirements for what you need to run it.

No llmrun Composite Score for Llama 3.1 405B: the score is only published for a model measured on at least 4 of the boards in the score panel, spanning at least 2 benchmark categories and including a reasoning or coding board — or on 3 boards that narrow its score to a 90% range of at most 15.6 points. Either way the results must come from at least two different benchmarks. Its individual scores below stand on their own — see how the llmrun Composite Score gets calculated.

Llama 3.1 405B reasoning benchmarks

Multi-step logic and problem-solving — can the model think through a hard problem rather than recall an answer.

BIG-Bench Hard
BIG-Bench Hard (BBH) is the 23-task subset of BIG-Bench that earlier models found hardest — multi-step reasoning, logic, and word problems. It's a long-standing standard for tracking reasoning ability across model generations.
82.9%
82.9%
median 50.8%open leader 87.5%
Overall best: Gemini 1.5 Pro 001 89.2%
97th percentile · #2 of 38 open · #4 of 50 all
#2 of 38
#4 of 50
Lab-reported
Epoch AI · Jul 2024

Other results (2)· leaderboard-only, not part of the llmrun Score

WinoGrande
WinoGrande is a large-scale commonsense-reasoning test built around pronoun resolution: the model must pick which of two nouns a pronoun refers to in a sentence, using real-world knowledge to disambiguate. It was designed by AI2 to resist the shortcuts that let earlier models game its predecessor, Winograd.
89.2%
89.2%
median 70.1%open leader
top · #1 of 51 open · #1 of 79 all
#1 of 51
#1 of 79
Lab-reported
Epoch AI · Jul 2024
PIQA
PIQA (Physical Interaction QA) asks a model to choose the more sensible of two ways to accomplish an everyday physical task, testing practical, physical-world commonsense that isn't well captured by purely textual reasoning tests.
85.9%
85.9%
median 80.1%open leader 88.6%
Overall best: GPT 4o Mini (Jul 18, 2024) 88.7%
97th percentile · #2 of 40 open · #4 of 59 all
#2 of 40
#4 of 59
Lab-reported
Epoch AI · Jul 2024

Llama 3.1 405B knowledge benchmarks

Breadth and depth of factual knowledge across many subjects, paired with the reasoning to apply it.

MMLU-Pro
MMLU-Pro is a harder, cleaned-up successor to MMLU with ten answer choices and more reasoning-heavy questions across 14 subjects, measuring broad knowledge and reasoning together.
61.6%
61.6%
median 52.1%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
62nd percentile · #56 of 145 open · #135 of 257 all
#56 of 145
#135 of 257
Independent
TIGER-Lab

Other results (4)· leaderboard-only, not part of the llmrun Score

MMLU
MMLU (Massive Multitask Language Understanding) spans 57 subjects from history to law to medicine as multiple-choice questions. It is the long-standing default for broad knowledge and remains the most widely-reported general benchmark.
84.4%
84.4%
median 62.5%open leader 87.2%
Overall best: GPT 4o (Nov 20, 2024) 88.1%
93rd percentile · #7 of 85 open · #14 of 136 all
#7 of 85
#14 of 136
Lab-reported
Epoch AI · Jul 2024
ARC Challenge
ARC Challenge is a set of grade-school science questions, curated by the Allen Institute for AI (AI2) to exclude ones that simple word-matching or retrieval can answer. It measures science knowledge paired with the reasoning needed to apply it.
95.3%
95.3%
median 53.9%open leader 95.3%
99th percentile · #2 of 56 open · #2 of 76 all
#2 of 56
#2 of 76
Lab-reported
Epoch AI · Jul 2024
HellaSwag
HellaSwag is a commonsense benchmark that asks a model to pick the most plausible continuation of an everyday situation. The wrong options are adversarially chosen to fool models, testing grounded commonsense understanding.
89.2%
89.2%
median 76.2%open leader
Overall best: GPT 4 (Mar 14) 95.3%
top · #1 of 47 open · #3 of 75 all
#1 of 47
#3 of 75
Lab-reported
Epoch AI · Jul 2024
TriviaQA
TriviaQA is a large set of trivia questions paired with evidence documents, testing a model's factual recall and reading comprehension across a wide range of everyday topics.
82.7%
82.7%
median 73.7%open leader 87.6%
89th percentile · #3 of 19 open · #11 of 39 all
#3 of 19
#11 of 39
Lab-reported
Epoch AI · Jul 2024

Scores aggregated from public benchmark sources (Epoch AI: 7 boards, TIGER-Lab: 1 board) — each linked from the benchmark pages. llmrun does not run these benchmarks. Percentiles and ranks are among open models in the llmrun catalogue; the median and leader marks refer to the same field.