Noometry

Model comparison

Kimi K2.5 vs Qwen1.5-72B

Kimi K2.5 is the stronger model overall, scoring 47.8 to 30.3 on the Noometry Index.

Last verified . 18 shared benchmarks.

Kimi K2.5 Moonshot AI

47.8

Rank #50 Confirmed

Qwen1.5-72B Alibaba (Qwen)

30.3

Rank #282 Confirmed

Summary

  • They share 18 benchmarks with published results for both. Kimi K2.5 scores higher in 8 categories and Qwen1.5-72B in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Kimi K2.5 leads 53.6 to 11.5.
  • The biggest single-benchmark swing is GPQA Diamond: 87.6% for Kimi K2.5 and 28.8% for Qwen1.5-72B.

Side by side

Kimi K2.5Qwen1.5-72B
ProviderMoonshot AIAlibaba (Qwen)
Noometry Index47.830.3
Released2026-01-272024-02-04
WeightsOpenOpen
Context window262K—
Max output262K—
Input $ / M tokens$0.45—
Output $ / M tokens$2.25—
Results tracked5122

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

Category by category

Coding Kimi K2.5 leads

Kimi K2.5: 48.7 (#48), Qwen1.5-72B: 31.4 (#272)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Coding14741166
SWE-bench Verified73.8%—
SWE-bench Verified (bash only)70.8%—
LMArena WebDev1436—
SWE-bench Multilingual67.3%—
SciCode49%—
WeirdML45.6%—
BigCodeBench Instruct—33.2%
BigCodeBench Complete—40.3%
ALE-Bench821.65—
HumanEval+—59.1%
MBPP+—61.6%

Agentic & Tool Use Not comparable

Kimi K2.5: 34.0 (#44), Qwen1.5-72B: —

BenchmarkKimi K2.5Qwen1.5-72B
Terminal-Bench43.2%—
OSWorld63.3%—
Vending-Bench 21,198—

Reasoning Kimi K2.5 leads

Kimi K2.5: 31.2 (#74), Qwen1.5-72B: 21.2 (#201)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Hard Prompts14541148
ARC-AGI-211.8%—
SimpleBench46.8%—
Kagi LLM Benchmark78.5%—
NYT Connections (extended)69.9%—
ARC-AGI-165.3%—
CritPt3.1%—
Chess Puzzles12%—
EnigmaEval3.4%—
Thematic Generalization69.4%—
Epoch Capabilities Index148.05—

Math Kimi K2.5 leads

Kimi K2.5: 50.3 (#46), Qwen1.5-72B: 30.7 (#205)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Math14701164
MathArena Final-Answer Competitions62.3%—
OTIS Mock AIME 2024-202592.2%—
FrontierMath (Feb 2025 set)27.9%—
FrontierMath Tier 4 (v1)4.2%—

Knowledge Kimi K2.5 leads

Kimi K2.5: 53.6 (#52), Qwen1.5-72B: 11.5 (#295)

BenchmarkKimi K2.5Qwen1.5-72B
GPQA Diamond87.6%28.8%
LMArena Expert14661136
Humanity's Last Exam24.4%—
SimpleQA Verified34.3%—
Vectara Hallucination Rate14.2%—

Multimodal Not comparable

Kimi K2.5: 40.4 (#37), Qwen1.5-72B: —

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Vision1269—
LMArena Document1430—

Multilingual Kimi K2.5 leads

Kimi K2.5: 54.1 (#50), Qwen1.5-72B: 33.4 (#248)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Non-English14331135
LMArena Chinese14951186
LMArena French14541158
LMArena German14411084
LMArena Japanese14221061
LMArena Korean14121049
LMArena Russian14361104
LMArena Spanish14501110

Instruction Following Kimi K2.5 leads

Kimi K2.5: 75.5 (#60), Qwen1.5-72B: 59.6 (#251)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Instruction Following14311142

Long Context Kimi K2.5 leads

Kimi K2.5: 52.1 (#7), Qwen1.5-72B: 35.0 (#238)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Longer Query14451157
Fiction.LiveBench86.1%—
CL-bench19.3%—
CL-bench Life13.2%—

Writing & Preference Kimi K2.5 leads

Kimi K2.5: 65.3 (#51), Qwen1.5-72B: 37.4 (#253)

BenchmarkKimi K2.5Qwen1.5-72B
LMArena Text14451166
LMArena Creative Writing14231137
LMArena Multi-Turn14441160
EQ-Bench Creative Writing1579—

Frequently asked questions

Is Kimi K2.5 better than Qwen1.5-72B?

Kimi K2.5 is the stronger model overall, scoring 47.8 to 30.3 on the Noometry Index.

Is Kimi K2.5 or Qwen1.5-72B better for coding?

Kimi K2.5 scores higher on coding benchmarks: 48.7 versus 31.4 in the Noometry coding category.

How many benchmarks do Kimi K2.5 and Qwen1.5-72B share?

18 benchmarks have published results for both models. Kimi K2.5 has 51 scored results on Noometry and Qwen1.5-72B has 22.

Related comparisons

Go deeper