Model comparison

Kimi K3 vs Step 3

Kimi K3 is the stronger model overall, scoring 59.5 to 40.5 on the Noometry Index.

Last verified . 15 shared benchmarks.

Kimi K3 Moonshot AI

59.5

Rank #15 Confirmed

Step 3 StepFun

40.5

Rank #149 Confirmed

Summary

  • They share 15 benchmarks with published results for both. Kimi K3 scores higher in 9 categories and Step 3 in 0 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in math, where Kimi K3 leads 74.2 to 37.6.

Side by side

Kimi K3 and Step 3 specifications
Kimi K3Step 3
ProviderMoonshot AIStepFun
Noometry Index59.540.5
Released2026-07-16—
WeightsOpenOpen
Context window1.05M—
Max output1.05M—
Input $ / M tokens$3—
Output $ / M tokens$15—
Results tracked5317

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

Category by category

Coding Kimi K3 leads

Kimi K3: 61.0 (#10), Step 3: 40.1 (#147)

Coding benchmarks
BenchmarkKimi K3Step 3
LMArena Coding15081367
DeepSWE68.5%—
FrontierCode44.2%—
LMArena WebDev1654—
FrontierSWE25.9%—
SciCode59.5%—
WeirdML82.6%—
ALE-Bench1,524—

Agentic & Tool Use Not comparable

Kimi K3: 41.8 (#20), Step 3: —

Agentic & Tool Use benchmarks
BenchmarkKimi K3Step 3
APEX-Agents50.6%—
τ²-bench Banking37.1%—
PostTrainBench32%—
GBAEval48.3%—
GDP.pdf19%—
Vending-Bench 25,165—

Reasoning Kimi K3 leads

Kimi K3: 63.0 (#17), Step 3: 28.4 (#105)

Reasoning benchmarks
BenchmarkKimi K3Step 3
LMArena Hard Prompts14961355
ARC-AGI-260.4%—
SimpleBench60.7%—
Kagi LLM Benchmark—62.3%
NYT Connections (extended)93.6%—
ARC-AGI-194.5%—
CritPt23.4%—
Chess Puzzles39%—
Mystery Game Puzzles26%—
DTBench91.2%—
LMCA52.7%—
Surface Evolver Bench95%—
Epoch Capabilities Index157.45—
ForecastBench61.1—

Math Kimi K3 leads

Kimi K3: 74.2 (#16), Step 3: 37.6 (#148)

Knowledge Kimi K3 leads

Kimi K3: 63.2 (#21), Step 3: 36.8 (#164)

Knowledge benchmarks
BenchmarkKimi K3Step 3
LMArena Expert15211333
GPQA Diamond93.1%—
SimpleQA Verified50.6%—

Multimodal Kimi K3 leads

Kimi K3: 37.8 (#70), Step 3: 35.5 (#86)

Multimodal benchmarks
BenchmarkKimi K3Step 3
LMArena Vision—1177
Blueprint-Bench 229.5%—
Furniture Assembly34.2%—

Multilingual Kimi K3 leads

Kimi K3: 56.3 (#21), Step 3: 46.3 (#159)

Multilingual benchmarks
BenchmarkKimi K3Step 3
LMArena Non-English14661327
LMArena Chinese15291397
LMArena German14881371
LMArena Korean14581269
LMArena Russian14821331
LMArena Spanish14721371
LMArena French1491—
LMArena Japanese1487—

Instruction Following Kimi K3 leads

Kimi K3: 77.7 (#14), Step 3: 70.4 (#164)

Instruction Following benchmarks
BenchmarkKimi K3Step 3
LMArena Instruction Following14831332

Long Context Kimi K3 leads

Kimi K3: 45.8 (#29), Step 3: 40.3 (#157)

Long Context benchmarks
BenchmarkKimi K3Step 3
LMArena Longer Query14941326

Writing & Preference Kimi K3 leads

Kimi K3: 76.6 (#4), Step 3: 54.3 (#151)

Writing & Preference benchmarks
BenchmarkKimi K3Step 3
LMArena Text14761350
LMArena Creative Writing14541321
LMArena Multi-Turn14881341
EQ-Bench Creative Writing2082—
EQ-Bench 41339—

Frequently asked questions

Is Kimi K3 better than Step 3?

Kimi K3 is the stronger model overall, scoring 59.5 to 40.5 on the Noometry Index.

Is Kimi K3 or Step 3 better for coding?

Kimi K3 scores higher on coding benchmarks: 61.0 versus 40.1 in the Noometry coding category.

How many benchmarks do Kimi K3 and Step 3 share?

15 benchmarks have published results for both models. Kimi K3 has 53 scored results on Noometry and Step 3 has 17.

Related comparisons

Go deeper