Noometry

Model comparison

Claude 3 Opus vs Llama 3.1-8B

Claude 3 Opus is the stronger model overall, scoring 29.5 to 23.0 on the Noometry Index.

Last verified . 30 shared benchmarks.

Claude 3 Opus Anthropic

29.5

Rank #310 Confirmed

Llama 3.1-8B Meta

23.0

Rank #352 Confirmed

Summary

  • They share 30 benchmarks with published results for both. Claude 3 Opus scores higher in 8 categories and Llama 3.1-8B in 1 category; 8 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Claude 3 Opus leads 47.2 to 29.7.
  • The biggest single-benchmark swing is GPQA Diamond: 47.2% for Claude 3 Opus and 27% for Llama 3.1-8B.
  • Llama 3.1-8B has downloadable open weights; the other is API-only.

Side by side

Claude 3 OpusLlama 3.1-8B
ProviderAnthropicMeta
Noometry Index29.523.0
Released2024-02-292024-07-23
WeightsProprietaryOpen
Context window—128K
Max output—4K
Input $ / M tokens—$0.05
Output $ / M tokens—$0.08
Results tracked4643

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Claude 3 Opus leads

Claude 3 Opus: 32.9 (#267), Llama 3.1-8B: 20.2 (#340)

BenchmarkClaude 3 OpusLlama 3.1-8B
WeirdML19.2%1.7%
BigCodeBench Instruct45.5%32.8%
LMArena Coding12641195
BigCodeBench Complete57.4%40.5%
HumanEval+77.4%62.8%
MBPP+73.3%55.6%
SciCode—13.2%
LiveBench Coding38.6%—

Agentic & Tool Use Claude 3 Opus leads

Claude 3 Opus: 24.6 (#116), Llama 3.1-8B: 22.5 (#131)

BenchmarkClaude 3 OpusLlama 3.1-8B
Berkeley Function Calling Leaderboard—25.8%
Cybench10%—
BALROG—15.1%
METR Time Horizons29.5%—

Reasoning Too close to call

Claude 3 Opus: 14.6 (#324), Llama 3.1-8B: 14.9 (#321)

BenchmarkClaude 3 OpusLlama 3.1-8B
Chess Puzzles5%0%
LMArena Hard Prompts12451175
DTBench61.6%50.9%
LMCA17%5.4%
Epoch Capabilities Index126.91116.57
SimpleBench23.5%—
CritPt—0%
EnigmaEval0.8%—
LiveBench Reasoning40.6%—
LiveBench Data Analysis57.9%—
ForecastBench58.4—
LiveBench49.2%—
PIQA—81.2%
WinoGrande88.5%—

Math Claude 3 Opus leads

Claude 3 Opus: 14.8 (#299), Llama 3.1-8B: 10.2 (#317)

BenchmarkClaude 3 OpusLlama 3.1-8B
OTIS Mock AIME 2024-20254.7%1.7%
LMArena Math12731179
MATH Level 537.5%22.9%
Omni-MATH—13.7%
LiveBench Math43.6%—
GSM8K—82.4%

Knowledge Claude 3 Opus leads

Claude 3 Opus: 24.5 (#267), Llama 3.1-8B: 8.0 (#307)

BenchmarkClaude 3 OpusLlama 3.1-8B
GPQA Diamond47.2%27%
LMArena Expert12231144
MMLU84.6%56.1%
SimpleQA Verified12.6%—
MMLU-Pro—40.6%
Confabulations22.7%—
GPQA (HELM)—24.7%
BoolQ—82.8%

Multimodal Not comparable

Claude 3 Opus: 27.1 (#116), Llama 3.1-8B: —

BenchmarkClaude 3 OpusLlama 3.1-8B
LMArena Vision1023—

Multilingual Claude 3 Opus leads

Claude 3 Opus: 41.4 (#207), Llama 3.1-8B: 34.0 (#249)

BenchmarkClaude 3 OpusLlama 3.1-8B
LMArena Non-English12581148
LMArena Chinese12481151
LMArena French12751177
LMArena German12581144
LMArena Japanese12041061
LMArena Korean11871053
LMArena Russian12801158
LMArena Spanish12461169

Instruction Following Claude 3 Opus leads

Claude 3 Opus: 64.1 (#228), Llama 3.1-8B: 58.9 (#258)

BenchmarkClaude 3 OpusLlama 3.1-8B
LMArena Instruction Following12481159
LiveBench Instruction Following63.9%—
IFEval—74.3%

Long Context Claude 3 Opus leads

Claude 3 Opus: 38.2 (#202), Llama 3.1-8B: 35.8 (#238)

BenchmarkClaude 3 OpusLlama 3.1-8B
LMArena Longer Query12591182

Writing & Preference Claude 3 Opus leads

Claude 3 Opus: 47.2 (#213), Llama 3.1-8B: 29.7 (#290)

BenchmarkClaude 3 OpusLlama 3.1-8B
LMArena Text12621187
LMArena Creative Writing12351154
LMArena Multi-Turn12751172
EQ-Bench Creative Writing—713
WildBench—68.7%
LiveBench Language50.4%—

Frequently asked questions

Is Claude 3 Opus better than Llama 3.1-8B?

Claude 3 Opus is the stronger model overall, scoring 29.5 to 23.0 on the Noometry Index.

Is Claude 3 Opus or Llama 3.1-8B better for coding?

Claude 3 Opus scores higher on coding benchmarks: 32.9 versus 20.2 in the Noometry coding category.

How many benchmarks do Claude 3 Opus and Llama 3.1-8B share?

30 benchmarks have published results for both models. Claude 3 Opus has 46 scored results on Noometry and Llama 3.1-8B has 43.

Related comparisons

Go deeper