Skip to content
CallMissed
Book a demo

Explore CallMissed

Back to all models
LLM Chatnewvisionreasoningtools
Moonshot — kimi-k3.1-pro logo

Kimi K3.1 Pro

by Moonshot · Released 2026

Moonshot AI's most capable open-weight model: natively multimodal and built for agentic work. 1M context window, image input, tool calling, JSON mode, and selectable reasoning.

Moonshot — kimi-k3.1-pro logo
LLM Chat

Kimi K3.1 Pro

Powered by Moonshot · Mixture-of-Experts (2.8T params, 104B active, agentic multimodal)

Context Window

1M

Parameters

2.8T (MoE, 104B active)

Max Output

131K

Category

LLM Chat

Overview

Kimi K3.1 Pro is Moonshot AI's most capable 3T-class model: an open-weight, natively multimodal agentic model post-trained from Kimi K3 with large-scale joint reinforcement learning. Moonshot reports improvements in coding, 3D creation, financial research, factual accuracy and agentic work.

It is a mixture-of-experts model with 2.8T total and 104B active parameters, 896 experts with 16 selected per token, Kimi Delta Attention and Gated MLA, and a MoonViT-V2 vision encoder. It takes text and image input and returns text, with a 1,048,576-token context window and 131,072 tokens of default max output.

Served through CallMissed, Kimi K3.1 Pro is OpenAI-compatible (`/v1/chat/completions`) with streaming, tool calling, JSON mode and a `reasoning_effort` control. It always thinks; send `"low"`, `"high"` or `"max"` to pick a level (`"low"` is fastest). It is available on paid plans only.

Pricing

MetricPrice
Input /1M tokens₹200.0000
Output /1M tokens₹1000.0000

1 credit = ₹1 ≈ US$0.0104 (US$1 = ₹96). Transparent per-call pricing — you pay only for what you call, with no per-seat charge or minimums.

Key Highlights

  • Moonshot's most capable open-weight agentic model
  • 1M-token context with 131K default max output
  • Image input, tool calling and JSON mode
  • Always-on reasoning at low, high or max

Technical Details

  • Mixture-of-experts: 2.8T total / 104B active parameters
  • 896 experts, 16 selected per token
  • Kimi Delta Attention + Gated MLA, MoonViT-V2 vision encoder
  • Context window: 1,048,576 tokens; default max output: 131,072 tokens
  • Input: text and image; output: text
  • Function/tool calling, JSON mode and streaming
  • reasoning_effort: "low", "high", "max"; "none" and "minimal" run as "low", "medium" as "high", "xhigh" as "max"
  • Cached input billed at $0.20 per 1M tokens
  • OpenAI-compatible on the CallMissed gateway; paid plans only

Strengths

  • 1M context keeps large codebases and documents in view
  • Image input, tool calling and JSON mode in one model
  • Selectable reasoning depth, including a full off switch

Limitations

  • Frontier-scale model, so higher latency than small fast tiers
  • Paid plans only; not part of the free tier
  • Listed language support is English

Use Cases

Agentic codingFinancial researchTool-using agentsImage-grounded analysis

API Example

curl https://api.callmissed.com/v1/chat/completions \
  -H "Authorization: Bearer cm_YOUR_KEY" \
  -d '{"model": "kimi-k3.1-pro", "messages": [{"role": "user", "content": "Plan and implement this feature across the repo"}]}'

Endpoint: POST /v1/chat/completions · Model ID: kimi-k3.1-pro

Try Kimi K3.1 Pro now

Get 1000 free API credits on signup. No credit card required.

Start free Book a DemoRead docs