LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Kimi K2 Instruct 0905

moonshotai·moonshotai/kimi-k2-0905·GA·Open weights·kimi-k2 series
September 2025 Kimi K2 instruction update with a 256K context and improved agentic coding

Specs & pricing

Input / output per 1M tokens
Official price
—
Lowest paid·NanoGPTGateway
$0.40 / $1.80
Blended $0.75 · 2.6× spread
Context
262,144
Output limit
98,304
Knowledge cutoff
2024-10
Released / updated
2025-09-05 / 2025-09-05
Capabilities
Reasoning✓ Tool use✓ Structured output✓ TemperatureAttachments
Modalities
Text
License
Modified MIT
Weights
Hugging Face

Kimi K2 Instruct 0905 by moonshotai is offered by 14 providers on this page. Public prices are shown for 13 of them. No official API price is listed for it; the prices below come from third-party channels. The lowest paid channel is NanoGPT at $0.40 / $1.80 per 1M. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. Prices vary about 2.6× across the providers that host it.

Median output speed is 44 tokens per second, with 1.8s to the first token.

The context window is 262,144 tokens at the reference host, but hosts report different limits, from 32,768 to 262,144, so the usable window depends on the provider you pick. It supports tool use and structured output. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Its training knowledge cuts off at 2024-10.

Quality & performanceArtificial Analysis · Intelligence Index v5

Intelligence—
Value—
Output speed44 tok/s
TTFT1.8 s

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 14 providers13 with public prices

  • ProviderInput / Output
  • NanoGPTGateway
    moonshotai/Kimi-K2-Instruct-0905
    $0.40 / $1.80
    Cache read $0.20Context 262,144Output limit 100,352
  • HeliconeGateway
    $0.50 / $2.00
    Cache read $0.40Context 262,144Output limit 16,384
  • OpenCodeGateway
    kimi-k2
    $0.40 / $2.50
    Cache read $0.40Context 262,144Output limit 262,144deprecated
  • IO.NETCloud
    moonshotai/Kimi-K2-Instruct-0905
    $0.57 / $2.30
    Cache read $0.29Context 32,768 ⚠Output limit 4,096
  • Jiekou.AIGateway
    $0.60 / $2.50
    Context 262,144Output limit 100,352
  • Novita AIGateway
    $0.60 / $2.50
    Context 262,144Output limit 98,304
  • OpenRouterGateway
    $0.60 / $2.50
    Context 262,144Output limit 98,304
  • Prime InferenceGateway
    $0.60 / $2.50
    Context 262,144Output limit 98,304
  • Kilo GatewayGateway
    $0.60 / $2.50
    Context 262,144Output limit 98,304
  • 302.AIGateway
    kimi-k2-0905-preview
    $0.63 / $2.51
    Context 262,144Output limit 262,144
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
NanoGPT
moonshotai/Kimi-K2-Instruct-0905
Gateway$0.40$1.80$0.20—262,144100,352
HeliconeGateway$0.50$2.00$0.40—262,14416,384
OpenCode
kimi-k2
Gateway$0.40$2.50$0.40—262,144262,144deprecated
IO.NET
moonshotai/Kimi-K2-Instruct-0905
Cloud$0.57$2.30$0.29—32,768 ⚠4,096
Jiekou.AIGateway$0.60$2.50——262,144100,352
Novita AIGateway$0.60$2.50——262,14498,304
OpenRouterGateway$0.60$2.50——262,14498,304
Prime InferenceGateway$0.60$2.50——262,14498,304
Kilo GatewayGateway$0.60$2.50——262,14498,304
302.AI
kimi-k2-0905-preview
Gateway$0.63$2.51——262,144262,144

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Your usage cost

Common scenarios
1NanoGPT$146.00
2Helicone$188.00
3IO.NET$194.80
4OpenCode$205.00
5Jiekou.AI$245.00
6Novita AI$245.00
The cheapest paid channel is the only channel.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$2.50$02026-08-052026-09-07
2026-09-07Input list —Output list —Min blended $0.75

Related models

Kimi K2.8 Previewsame series—Kimi K2.7 Codesame series$0.95 / $4.00Kimi K2.7 Code Highspeedsame series—Kimi K2 0711same lab—Kimi K2 Instructsame lab—
Data partly from models.dev (MIT)