Alla modellerchatKimi K3

Kimi K3

av MoonshotPremium

Testa på Kunya
chat

Kimi K3

2.8T MoE flagship for complex coding, knowledge work, and long-horizon agentic workflows with 1M context and native vision

Modell-ID:kimi-k3

API-dokumentation

Generate chat completions with this model.

POST/api/v1/chat/completions
GET /api/v1/models?type=chat
Request body
{
  "model": "kimi-k3",
  "messages": [
    {
      "role": "user",
      "content": "Hello!"
    }
  ],
  "stream": false,
  "thinking": true
}
cURL-exempel
curl -X POST https://kunya.ai/api/v1/chat/completions \
  -H "Authorization: Bearer kunya_YOUR_API_KEY" \
  -H "Content-Type: application/json"
  -d '{"model":"kimi-k3","messages":[{"role":"user","content":"Hello!"}],"stream":false,"thinking":true}'

Anteckningar

  • •OpenAI-compatible endpoint — works with any Kunya API key.
  • •Set `"stream": true` for server-sent events.
  • •Supports vision: pass image URLs in message content as `{ "type": "image_url", "image_url": { "url": "..." } }`.
  • •Supports tool calling via the `tools` parameter.

Priser

Indata$3.9 per 1M tokens
Utdata$19.5 per 1M tokens
Kontextfönster1049K

Kunya benchmark

Updated 2026-10-06 · #8 performance · #20 value

Performance
95.9
Value (KVI)
60.9
Literary writing
81.3
SWE Verified
93.4%
Cost
$18.47/1k tasks
Full benchmark report →

Funktioner

Streaming Ja
Vision Ja
Resonemang Ja
Verktyg Ja
LeverantörMoonshot
Testa på Kunya

Liknande modeller

Kimi K2.7 Code

Moonshot

Coding-focused MoE model for long-horizon programming, agentic task decomposition, and multimodal reasoning

API-dokumentation

Kimi K2.5

Moonshot

State-of-the-art visual coding and agentic tool-calling with multimodal reasoning

API-dokumentation

Claude Opus 4.6

Anthropic

Hybrid reasoning model with 1M context, top-tier coding and agentic performance

API-dokumentation

Claude Sonnet 5

Anthropic

Previous Sonnet — prefer Claude Sonnet 5.5 for speed and cost per task

API-dokumentation