Noometry

Model comparison

Kimi K2 Thinking Turbo vs Qwen2.5-Coder-32B

Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.4 to 32.9 on the Noometry Index.

Last verified . 12 shared benchmarks.

Kimi K2 Thinking Turbo Moonshot AI

45.4

Rank #63 Confirmed

Qwen2.5-Coder-32B Alibaba (Qwen)

32.9

Rank #239 Confirmed

Summary

  • They share 12 benchmarks with published results for both. Kimi K2 Thinking Turbo scores higher in 8 categories and Qwen2.5-Coder-32B in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in writing & preference, where Kimi K2 Thinking Turbo leads 60.2 to 41.8.

Side by side

Kimi K2 Thinking TurboQwen2.5-Coder-32B
ProviderMoonshot AIAlibaba (Qwen)
Noometry Index45.432.9
Released2025-11-062024-09-18
WeightsOpenOpen
Context window—33K
Max output—29K
Input $ / M tokens—$0.66
Output $ / M tokens—$1
Results tracked2131

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

Category by category

Coding Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 38.1 (#166), Qwen2.5-Coder-32B: 22.4 (#328)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Coding14541276
SWE-bench Verified (bash only)—9%
Aider Polyglot—16.4%
LMArena WebDev1322—
BigCodeBench Instruct—49%
LiveBench Coding—56.9%
BigCodeBench Complete—58%
HumanEval+—87.2%
MBPP+—77%

Reasoning Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 31.4 (#72), Qwen2.5-Coder-32B: 20.4 (#220)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Hard Prompts14281251
Chess Puzzles20%—
LiveBench Reasoning—42.1%
LiveBench Data Analysis—49.9%
Epoch Capabilities Index—119.41
HellaSwag—83%
LiveBench—46.2%
WinoGrande—80.8%

Math Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 45.5 (#62), Qwen2.5-Coder-32B: 30.8 (#203)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Math14291251
OTIS Mock AIME 2024-202583.1%—
LiveBench Math—46.6%
GSM8K—93%

Knowledge Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 50.9 (#64), Qwen2.5-Coder-32B: 32.9 (#200)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Expert14391221
GPQA Diamond84.2%—
ARC (AI2) Challenge—70.5%
MMLU—79.1%

Multilingual Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 51.6 (#104), Qwen2.5-Coder-32B: 38.0 (#230)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Non-English13981205
LMArena Chinese14571222
LMArena Russian13901228
LMArena French1425—
LMArena German1390—
LMArena Japanese1356—
LMArena Korean1329—
LMArena Spanish1407—

Instruction Following Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 74.2 (#103), Qwen2.5-Coder-32B: 61.6 (#240)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Instruction Following14031223
LiveBench Instruction Following—58.7%

Long Context Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 43.2 (#99), Qwen2.5-Coder-32B: 37.9 (#204)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Longer Query14151251

Writing & Preference Kimi K2 Thinking Turbo leads

Kimi K2 Thinking Turbo: 60.2 (#99), Qwen2.5-Coder-32B: 41.8 (#235)

BenchmarkKimi K2 Thinking TurboQwen2.5-Coder-32B
LMArena Text14151230
LMArena Creative Writing13731174
LMArena Multi-Turn14141222
LiveBench Language—23.3%

Frequently asked questions

Is Kimi K2 Thinking Turbo better than Qwen2.5-Coder-32B?

Kimi K2 Thinking Turbo is the stronger model overall, scoring 45.4 to 32.9 on the Noometry Index.

Is Kimi K2 Thinking Turbo or Qwen2.5-Coder-32B better for coding?

Kimi K2 Thinking Turbo scores higher on coding benchmarks: 38.1 versus 22.4 in the Noometry coding category.

How many benchmarks do Kimi K2 Thinking Turbo and Qwen2.5-Coder-32B share?

12 benchmarks have published results for both models. Kimi K2 Thinking Turbo has 21 scored results on Noometry and Qwen2.5-Coder-32B has 31.

Related comparisons

Go deeper