Skip to content
All systems operational0 AI providers monitored, polled every 2 minutes
Live status

Kimi K3

Flagship

by Moonshot AI

Kimi K3 is Moonshot AI's flagship, released July 16, 2026 and the largest open-weight system anyone has shipped: 2.8 trillion total parameters across 896 experts with roughly 16 active per token, so only about 1.8 percent of the network fires on any given token. It carries a 1 million token context window and native vision. API pricing is $0.30 per million cache-hit input tokens, $3.00 on a cache miss, and $15.00 output. Moonshot reports roughly 57 on the Artificial Analysis Intelligence Index, placing it third or fourth overall behind Claude Fable 5 and GPT-5.6 Sol, alongside first place on Frontend Code Arena at 1,679 points, the strongest open-weight GPQA Diamond result on record at 93.5 percent, and third on GDPval-AA v2 at 1,687; all vendor-reported. On normalized cost per task it is the cheapest of the leaders. Downloadable is not the same as runnable here: it ships natively in MXFP4 with no FP16 checkpoint, and even at 4-bit it needs roughly 1,450GB of VRAM, which means eight B200-class cards. This needs a rack, not a workstation.

Input Price

$3.00

per 1M tokens

Output Price

$15.00

per 1M tokens

Context Window

1M

tokens

Released

2026-07

Open source

Capabilities

textvisioncodetool-usereasoning

Key Strengths

  • Largest open-weight model ever shipped at 2.8T parameters
  • Sparse MoE fires only about 1.8 percent of the network per token
  • 1M context with native vision
  • Vendor-reported 93.5 on GPQA Diamond, best open-weight result on record
  • Cheapest normalized cost per task among the index leaders

Best For

  • Frontier-class self-hosting at rack scale
  • Frontend and UI code generation
  • Long-context multimodal research
  • Sovereign deployments needing open weights

Benchmark Scores

BenchmarkScoreDescription
GPQA Diamond93.5Graduate-level science questions verified by domain experts

Scores sourced from public benchmark datasets. See full benchmark leaderboard for all models.

Pricing Details

Input tokens

$3.00

per 1M tokens

Output tokens

$15.00

per 1M tokens

Estimated cost per 1K requests

$10.50

~1K input + ~500 output tokens avg

Prices are subject to change. Check the official documentation for current pricing. See the cost calculator for detailed estimates.

Open Source Model

Kimi K3 is free to download and self-host under the Modified MIT. Hosted API pricing varies by provider (e.g., Together, Fireworks, Groq). See our open source LLM guide for deployment options.

Related Models

View DocumentationCompare ModelsCost CalculatorFull Pricing Guide