Model comparison

Kimi K2 0905 vs GPT-5

Capability radar, value bars, and side-by-side posture.

← Back to Compare

Kimi K2 0905
Not scored
value score
GPT-5
94
value score
Kimi K2 0905 ctx
GPT-5 ctx
400K

Capability radar

Value · context · multimodal · openness · speed posture

Value head-to-head

  • GPT-594

Attribute tape

FieldKimi K2 0905GPT-5
Developer~moonshotaiOpenAI
Context400K
ModalitiesText, Code, Image, Audio
OpennessProprietary Model
SpeedFast
PriceMid-High
Value score94
SummaryKimi K2 0905 is an updated Mixture-of-Experts large language model from Moonshot AI with 1T total parameters and 32B active per forward pass, optimized for agentic coding, tool use, and long-context reasoning up to 256k tokens. It is available as an open-weight instruct model and via hosted API surfaces including OpenRouter.OpenAI frontier multimodal model.
Open Kimi K2 0905Open GPT-5