← All models
Compare

Kimi K2 vs Kimi K3

Kimi K2 (Moonshot) and Kimi K3 (Moonshot) compared on benchmarks, pricing, context window, and use-case rankings.

AttributeKimi K2Moonshot · 17 in the newsKimi K3Moonshot · 22 in the news
Scores
Intelligence (ECI)140157
Coding999
Math585
Reasoning & Knowledge1580
Agentic & Tools2186
Specifications
DeveloperMoonshotMoonshot
FamilyKimiKimi
ReleasedSep 5, 2025Jul 16, 2026
Parameters1T2.8T
AvailabilityOpen weights (restricted use)Open weights (non-commercial)
Context window131K1M
Price — $/M input$0.57$3.00
Price — $/M output$2.30$14.00
Inputstexttext, image
Outputstexttext
Benchmarks
AIME 2024/202557%97%
GPQA Diamond77%93%
GSO (code optimization)5%
Humanity's Last Exam7%44%
LiveCodeBench56%
MMLU-Pro82%
SciCode35%59%
SimpleBench26%
Terminal-Bench28%
WeirdML39%83%
τ²-bench61%
APEX39%
ARC-AGI95%
ARC-AGI-260%
SimpleQA Verified43%
WebDev Arena1682

Use-case scores are 0–100 percentile composites across each area’s benchmarks, ranked against every model from the past year. Highlighted cells lead each row. Open a model for the full picture.

Frequently asked questions

Is Kimi K2 better than Kimi K3?

On Epoch AI's Capabilities Index, Kimi K3 scores higher (157) than Kimi K2 (140). The right pick depends on your task — compare their coding, math, and reasoning scores in the table above.

Which is cheaper, Kimi K2 or Kimi K3?

Kimi K2 is cheaper on input tokens at $0.57 per million, versus $3.00 (representative OpenRouter pricing).

Which has a larger context window, Kimi K2 or Kimi K3?

Kimi K3 supports up to 1M tokens, compared with 131K for the other.

Which is better for coding, Kimi K2 or Kimi K3?

Across coding benchmarks like SWE-bench Verified and Terminal-Bench, Kimi K3 ranks higher — 99th vs 9th percentile among the models tracked on Model Beat.

Want a different match-up? Open the compare tool to add or swap models.

Benchmarks & model data from Epoch AI (CC BY); pricing & specs from OpenRouter. ECI = Epoch Capabilities Index.