Writing

Kimi

Moonshot AI · #1 most active of 26 in Writing

Compare →

Moonshot AI's assistant, known for very long context windows and the open-weight K2 agentic model.

Current version

K3

Entry cost

Free consumer tier; per-token API (kimi-k3, k2.7-code, k2.6). K2.5 and moonshot-v1 sunset Aug 31, 2026.

Changes / 30d

5

Capabilities

  • Native multimodal (visual + text)
  • 1M-token context
  • Agentic tool use
  • Thinking / non-thinking modes
  • Dedicated coding models (k2.7-code, high-speed)
  • Open weights

Pricing

Free consumer tier; per-token API (kimi-k3, k2.7-code, k2.6). K2.5 and moonshot-v1 sunset Aug 31, 2026.

Entry price over time

7/26/2026 · $08/2/2026 · $0

Geek mode

Models under the hood

  • Kimi K3

    Moonshot AI

    disclosed
  • Kimi K2.7 Code

    Moonshot AI

    disclosed
  • Kimi K2.6

    Moonshot AI

    disclosed

Context window

1M tokens

Public API

yes

Multi-model routing

no

Stack signals

  • 2.8T parameters
  • Open-weight release on GitHub
  • Native visual understanding

Shared model stack

Other tracked products running on the same foundation models — a quick read on how much of the catalog moves when one of these models changes.

  • Kimi K2.6Moonshot AI · 1 product
  • Kimi K2.7 CodeMoonshot AI · 1 product
  • Kimi K3Moonshot AI · 1 product

green disclosed · amber inferred · grey unknown

Reported scale

Reference data. Each figure is whatever the source actually said — weekly users, downloads, revenue run-rate — with its own definition and date. These are not comparable between tools and are never used to rank anything.

Also in Writing

All Writing
Claude

Anthropic

2

Assistant for long-form and analysis

Version
Fable 5
Cost
from $20/mo
Model
Claude Fable 5
DeepSeek

DeepSeekChina

2

Open-weight reasoning models at a fraction of frontier pricing.

Version
V3.2
Cost
Free chat, API from $0.28/M input
Model
DeepSeek-V3.2
Qwen Chat

AlibabaChina

2

Alibaba's assistant on the widely forked Qwen model family.

Version
Qwen3-Max
Cost
Free tier, Model Studio pay-as-you-go
Model
Qwen3-Max
Z.ai GLM

Zhipu AIChina

2

GLM assistant with strong open-weight releases.

Version
GLM-4.7
Cost
Free tier, coding plan from $3/mo
Model
GLM-4.7
ChatGPT

OpenAI

1

General-purpose assistant

Version
GPT-5.2
Cost
Free tier, Plus $20/mo
Model
GPT-5.2
Doubao

ByteDanceChina

1

China's largest consumer AI assistant by monthly users.

Version
Doubao 1.8
Cost
Free consumer app, Volcano Engine per-token
Model
Doubao 1.8 Pro

Change history

  • model

    Model lineup now K3, K2.7 Code (and high-speed variant) and K2.6; K2.5 and moonshot-v1 sunset for new users on Aug 31, 2026.

    Kimi K2.5, Kimi K2 ThinkingKimi K3, Kimi K2.7 Code, Kimi K2.6

    source
  • version

    K2.5 ships with stronger agentic tool use and vision

    K2K2.5

    source
  • version

    Moonshot AI released Kimi K3, a 2.8T-parameter open-weight native multimodal agentic model.

    K2.5K3

    source
  • capability

    Context window expanded from 256K to 1M tokens with K3.

    256K tokens1M tokens

    source
  • capability

    Kimi Researcher moved out of invite-only preview

    Agentic tool useAgentic tool use, Kimi Researcher

    source