Noometry

Model comparison

Claude 3 Haiku vs Llama 3.1-405B

Llama 3.1-405B is the stronger model overall, scoring 30.4 to 25.7 on the Noometry Index.

Last verified . 29 shared benchmarks.

Claude 3 Haiku Anthropic

25.7

Rank #333 Confirmed

Llama 3.1-405B Meta

30.4

Rank #279 Confirmed

Summary

  • They share 29 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and Llama 3.1-405B in 8 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Llama 3.1-405B leads 30.4 to 17.3.
  • The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 49.8% for Llama 3.1-405B.
  • Llama 3.1-405B has downloadable open weights; the other is API-only.

Side by side

Claude 3 HaikuLlama 3.1-405B
ProviderAnthropicMeta
Noometry Index25.730.4
Released2024-03-072024-07-23
WeightsProprietaryOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked3742

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

Category by category

Coding Llama 3.1-405B leads

Claude 3 Haiku: 26.1 (#320), Llama 3.1-405B: 32.7 (#257)

BenchmarkClaude 3 HaikuLlama 3.1-405B
WeirdML9.8%21.4%
LMArena Coding11991291
BigCodeBench Instruct39.4%—
BigCodeBench Complete50.1%—
CadEval12%—
HumanEval+68.9%—
MBPP+68.8%—

Agentic & Tool Use Not comparable

Claude 3 Haiku: —, Llama 3.1-405B: 20.9 (#131)

BenchmarkClaude 3 HaikuLlama 3.1-405B
TheAgentCompany—7.4%
Cybench—7.5%

Reasoning Too close to call

Claude 3 Haiku: 15.7 (#303), Llama 3.1-405B: 16.4 (#293)

BenchmarkClaude 3 HaikuLlama 3.1-405B
Kagi LLM Benchmark34.2%45%
LMArena Hard Prompts11741270
DTBench50.1%61.4%
Epoch Capabilities Index118.3128.76
ForecastBench53.259.9
WinoGrande74.2%89.2%
SimpleBench—23%
LMCA8.8%—
BIG-Bench Hard—82.9%
HellaSwag—89.2%
PIQA—85.9%

Math Llama 3.1-405B leads

Claude 3 Haiku: 9.2 (#313), Llama 3.1-405B: 17.3 (#283)

BenchmarkClaude 3 HaikuLlama 3.1-405B
OTIS Mock AIME 2024-20251.8%9.7%
LMArena Math11881281
MATH Level 514.9%49.8%
Omni-MATH—24.9%

Knowledge Llama 3.1-405B leads

Claude 3 Haiku: 17.3 (#280), Llama 3.1-405B: 30.4 (#221)

BenchmarkClaude 3 HaikuLlama 3.1-405B
GPQA Diamond36.3%50.9%
Confabulations34.2%17.6%
LMArena Expert11481243
MMLU73.8%84.5%
MMLU-Pro—72.3%
GPQA (HELM)—52.2%
ARC (AI2) Challenge—95.3%
TriviaQA—82.7%

Multimodal Not comparable

Claude 3 Haiku: 23.1 (#123), Llama 3.1-405B: —

BenchmarkClaude 3 HaikuLlama 3.1-405B
LMArena Vision950—
ScienceQA72%—

Multilingual Llama 3.1-405B leads

Claude 3 Haiku: 36.1 (#238), Llama 3.1-405B: 40.9 (#209)

BenchmarkClaude 3 HaikuLlama 3.1-405B
LMArena Non-English11781248
LMArena Chinese11551242
LMArena French11941278
LMArena German11741252
LMArena Japanese11021208
LMArena Korean11081184
LMArena Russian12041265
LMArena Spanish11651259

Instruction Following Llama 3.1-405B leads

Claude 3 Haiku: 61.5 (#242), Llama 3.1-405B: 66.1 (#209)

BenchmarkClaude 3 HaikuLlama 3.1-405B
LMArena Instruction Following11741259
IFEval—81.1%

Long Context Llama 3.1-405B leads

Claude 3 Haiku: 36.1 (#232), Llama 3.1-405B: 38.4 (#192)

BenchmarkClaude 3 HaikuLlama 3.1-405B
LMArena Longer Query11911266

Writing & Preference Llama 3.1-405B leads

Claude 3 Haiku: 29.8 (#286), Llama 3.1-405B: 39.1 (#246)

BenchmarkClaude 3 HaikuLlama 3.1-405B
LMArena Text11951284
LMArena Creative Writing11571262
EQ-Bench Creative Writing717870
LMArena Multi-Turn11901297
WildBench—78.3%

Frequently asked questions

Is Claude 3 Haiku better than Llama 3.1-405B?

Llama 3.1-405B is the stronger model overall, scoring 30.4 to 25.7 on the Noometry Index.

Is Claude 3 Haiku or Llama 3.1-405B better for coding?

Llama 3.1-405B scores higher on coding benchmarks: 32.7 versus 26.1 in the Noometry coding category.

How many benchmarks do Claude 3 Haiku and Llama 3.1-405B share?

29 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Llama 3.1-405B has 42.

Related comparisons

Go deeper