Noometry

Model comparison

Claude 3 Haiku vs Qwen2-72B

Qwen2-72B is the stronger model overall, scoring 30.0 to 25.9 on the Noometry Index.

Last verified . 24 shared benchmarks.

Claude 3 Haiku Anthropic

25.9

Rank #340 Confirmed

Qwen2-72B Alibaba (Qwen)

30.0

Rank #300 Confirmed

Summary

  • They share 24 benchmarks with published results for both. Claude 3 Haiku scores higher in 1 category and Qwen2-72B in 7 categories; 5 gaps are clear of the uncertainty.
  • The widest gap is in math, where Qwen2-72B leads 30.2 to 9.8.
  • The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 39.1% for Qwen2-72B.
  • Qwen2-72B has downloadable open weights; the other is API-only.

Side by side

Claude 3 HaikuQwen2-72B
ProviderAnthropicAlibaba (Qwen)
Noometry Index25.930.0
Released2024-03-072024-06-07
WeightsProprietaryOpen
Context window——
Max output——
Input $ / M tokens——
Output $ / M tokens——
Results tracked3726

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

Category by category

Coding Qwen2-72B leads

Claude 3 Haiku: 26.4 (#325), Qwen2-72B: 29.1 (#310)

BenchmarkClaude 3 HaikuQwen2-72B
WeirdML9.8%11.3%
BigCodeBench Instruct39.4%38.5%
LMArena Coding11991196
BigCodeBench Complete50.1%54%
CadEval12%—
HumanEval+68.9%—
MBPP+68.8%—

Agentic & Tool Use Not comparable

Claude 3 Haiku: —, Qwen2-72B: 17.0 (#146)

BenchmarkClaude 3 HaikuQwen2-72B
TheAgentCompany—1.1%
METR Time Horizons—29.9%

Reasoning Qwen2-72B leads

Claude 3 Haiku: 16.3 (#307), Qwen2-72B: 23.2 (#181)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Hard Prompts11741191
Epoch Capabilities Index118.35125.28
Kagi LLM Benchmark34.2%—
DTBench50.1%—
LMCA8.8%—
ForecastBench53.2—
WinoGrande74.2%—

Math Qwen2-72B leads

Claude 3 Haiku: 9.8 (#319), Qwen2-72B: 30.2 (#236)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Math11881235
MATH Level 514.9%39.1%
OTIS Mock AIME 2024-20251.8%—

Knowledge Qwen2-72B leads

Claude 3 Haiku: 17.3 (#285), Qwen2-72B: 21.2 (#275)

BenchmarkClaude 3 HaikuQwen2-72B
GPQA Diamond36.3%40.8%
LMArena Expert11481171
MMLU73.8%82.4%
Confabulations34.2%—

Multimodal Not comparable

Claude 3 Haiku: 23.6 (#128), Qwen2-72B: —

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Vision950—
ScienceQA72%—

Multilingual Too close to call

Claude 3 Haiku: 36.0 (#243), Qwen2-72B: 35.9 (#244)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Non-English11781176
LMArena Chinese11551240
LMArena French11951170
LMArena German11741151
LMArena Japanese11021111
LMArena Korean11091083
LMArena Russian12041169
LMArena Spanish11661169

Instruction Following Too close to call

Claude 3 Haiku: 61.3 (#247), Qwen2-72B: 61.7 (#241)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Instruction Following11731181

Long Context Too close to call

Claude 3 Haiku: 36.1 (#237), Qwen2-72B: 36.1 (#235)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Longer Query11901192

Writing & Preference Qwen2-72B leads

Claude 3 Haiku: 29.7 (#291), Qwen2-72B: 40.8 (#241)

BenchmarkClaude 3 HaikuQwen2-72B
LMArena Text11951203
LMArena Creative Writing11571181
LMArena Multi-Turn11901196
EQ-Bench Creative Writing717—

Frequently asked questions

Is Claude 3 Haiku better than Qwen2-72B?

Qwen2-72B is the stronger model overall, scoring 30.0 to 25.9 on the Noometry Index.

Is Claude 3 Haiku or Qwen2-72B better for coding?

Qwen2-72B scores higher on coding benchmarks: 29.1 versus 26.4 in the Noometry coding category.

How many benchmarks do Claude 3 Haiku and Qwen2-72B share?

24 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Qwen2-72B has 26.

Related comparisons

Go deeper