Llama 3.1 70B Instruct — Benchmarks

Benchmark scores for Llama 3.1 70B Instruct aggregated from public leaderboards (10 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.

llmrun Composite Score →

77.9

90% range 73.2–82.7 · from 7 boards

CodingNot enough data yet
AgentsNot enough data yet
MathNot enough data yet
ScienceNot enough data yet
Reasoning83.290% range 74–93

Llama 3.1 70B Instruct reasoning benchmarks

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

GPQA Diamond
GPQA Diamond is a set of extremely hard, graduate-level science questions (physics, chemistry, biology) written by domain experts and filtered so that skilled non-experts with web access still fail. It measures genuine reasoning rather than memorization.
44.2%
44.2%
median 54.0%open leader 93.1%
Overall best: Gemini 4 Argon (high) 96.4%
33rd percentile · #66 of 98 open · #247 of 300 all
#66 of 98
#247 of 300
Independent
Epoch AI · Jul 2024
BALROG
BALROG (Benchmarking Agentic LLM and VLM Reasoning On Games) tests long-horizon agentic reasoning across six game environments — BabyAI, Crafter, TextWorld, Baba Is AI, MiniHack and NetHack — scoring the average progress a model makes toward completing each game. It was built by UCL's DARK lab and collaborators.
27.9%
27.9%
median 16.2%open leader 34.9%
Overall best: GPT 6 Astra Max 68.3%
93rd percentile · #2 of 15 open · #23 of 40 all
#2 of 15
#23 of 40
Independent
Epoch AI · Jul 2024
DTBench
DTBench is a set of handcrafted, expert-validated multiple-choice questions on decision theory — Newcomb-like problems and related puzzles — built by the Conceptual Reasoning Index project with Anthropic. It tests whether a model can follow an abstract argument rather than recall a fact.
60.0%
60.0%
median 67.7%open leader 93.9%
Overall best: Claude Opus 5.5 Max 98.9%
30th percentile · #48 of 68 open · #171 of 214 all
#48 of 68
#171 of 214
Lab-reported
Epoch AI · Jul 2024
LMCA
LMCA (Language Model Conceptual Argumentation) scores models on expert-rated arguments across philosophy, decision theory, and AI risk — domains where there is no empirical feedback loop to check an answer against. It is part of the Conceptual Reasoning Index, built with Anthropic by researchers Emery Cooper and Caspar Oesterheld.
14.8%
14.8%
median 24.7%open leader 55.5%
Overall best: Claude Opus 5.5 Max 68.2%
23rd percentile · #42 of 54 open · #156 of 176 all
#42 of 54
#156 of 176
Lab-reported
Epoch AI · Jul 2024

Llama 3.1 70B Instruct coding benchmarks

Writing, editing, and fixing real code — measured by whether the resulting program actually runs and passes tests.

WeirdML
WeirdML asks models to write machine-learning code for small, deliberately unusual datasets, then runs the code and scores the accuracy it actually achieves. Built by Håvard Tveit Ihle, it rewards practical problem-solving over recognizing a familiar task.
9.0%
9.0%
median 39.5%open leader 82.6%
Overall best: GPT 6 Astra Promax 93.6%
4th percentile · #44 of 46 open · #158 of 162 all
#44 of 46
#158 of 162
Independent
Epoch AI · Jul 2024

Llama 3.1 70B Instruct math benchmarks

Symbolic and quantitative problem-solving, from competition math to multi-step calculation.

AIME 2024/2025
AIME (American Invitational Mathematics Examination) is a prestigious high-school competition of hard, integer-answer problems. It's a widely-cited yardstick for multi-step mathematical reasoning, here on the 2024–2025 papers.
3.6%
3.6%
median 52.8%open leader 98.6%
Overall best: Claude Fable 5 (high) 100.0%
18th percentile · #71 of 86 open · #259 of 283 all
#71 of 86
#259 of 283
Independent
Epoch AI · Jul 2024
MATH Level 5
MATH Level 5 covers the hardest tier of competition-mathematics problems, testing multi-step symbolic and quantitative reasoning.
36.7%
36.7%
median 41.6%open leader 96.6%
Overall best: GPT 5 (Aug 07, 2025, high) 98.1%
42nd percentile · #23 of 39 open · #77 of 107 all
#23 of 39
#77 of 107
Independent
Epoch AI · Jul 2024

Llama 3.1 70B Instruct 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.
62.8%
62.8%
median 52.1%open leader 88.0%
Overall best: Gemini-3.1-Pro 91.2%
64th percentile · #53 of 145 open · #129 of 257 all
#53 of 145
#129 of 257
Independent
TIGER-Lab

Other results (1)· 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.
80.1%
80.1%
median 62.5%open leader 87.2%
Overall best: GPT 4o (Nov 20, 2024) 88.1%
89th percentile · #10 of 85 open · #24 of 136 all
#10 of 85
#24 of 136
Lab-reported
Epoch AI · Jul 2024

Llama 3.1 70B Instruct instruction following benchmarks

How precisely a model obeys explicit, verifiable instructions — format, length, and content constraints — and, more broadly, how much people prefer its responses head-to-head in blind human voting.

LMArena Text
Arena (arena.ai, formerly LMArena / Chatbot Arena) ranks models by millions of blind, head-to-head human votes on real prompts. The Arena Score shown here is a Bradley-Terry rating computed with Arena's default style control, so higher means people preferred that model's answers more often — it's a relative rating, not a percentage, so compare it only to other models on this board.
1293.4
1293.4
median 1321.5open leader 1487.1
Overall best: Gemini 4 Argon (High) 1523.9
42nd percentile · #109 of 187 open · #279 of 405 all
#109 of 187
#279 of 405
Arena
Arena · Oct 2026

Score vs VRAM in the ≤64 GB tier

Llama 3.1 70B Instruct
Other models
Best Score for its VRAM
x: VRAM (GB, log scale) · y: llmrun Score

#37 of 49 that fit ≤64 GB

Each dot is a model at the best quant that fits ≤64 GB. The line joins the models with the best Score for their VRAM; a dot below it is beaten by a model that needs no more memory.

Full ≤64 GB ranking →

Llama 3.1 70B Instruct vs its neighbours

Same boards for the nearest open models per llmrun Score. The highest score in each row is marked.

Board	Llama 3.1 70B Instructthis model	Qwen3.8 27Bbest in classQ4_K_M · 18 GB	C4ai Command A 03 2025nearest by ScoreQ4_K_M · 79 GB	Llama 3.3 70B Instructnearest by ScoreQ4_K_M · 51 GB	Qwen2 72B Instructnearest by ScoreQ4_K_M · 49 GB
llmrun Score	77.9	122.1	79.8	79.7	77.5
GPQA Diamond	44.2%	—	—	47.4%	40.8%
BALROG	27.9%	—	—	23.0%	—
DTBench	60.0%	88.0%	61.3%	59.5%	—
LMCA	14.8%	41.4%	10.3%	17.5%	—
WeirdML	9.0%	—	—	14.4%	11.3%
AIME 2024/2025	3.6%	—	—	5.1%	—
MATH Level 5	36.7%	—	—	41.6%	39.1%
MMLU-Pro	62.8%	—	—	65.9%	64.4%
LMArena Text	1293.4	1437.8	1353.7	1318.1	1261.9
Compare head to head		Compare	Compare	Compare	Compare

Scores aggregated from public benchmark sources (Epoch AI: 8 boards, Arena: 1 board, 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.