Alex Getmanearly AI signals
EN

Kimi K2.6

Moonshot

ReleasedUpdated
Context window
262K
262,144 tokens
Input / Output
image, text, video → text

API pricing

USD / 1M tokens

InputOutputCached input
  • $0.65
openrouter
  • $3.41
openrouter
  • $0.15
openrouter
Calculate your workload

Timeline

  1. · Weights version · evidence link
Input / output price · USD / 1M tokens
ModelEvidence linksInput / output price
moonshotai/kimi-k2.6
standard
openrouter$0.65 / $3.41
moonshotai/kimi-k2.6
standard
vercel-gateway$0.95 / $4.00
  1. · Evidence links · GitHub search · Moonshot · commits · vllm-project/vllm · @yiliu30
    Excerpt
    [Model] Load q/k/v scales for the quantized Kimi-K2.6 (#63097)
  2. · Evidence links · GitHub search · Moonshot · pulls · vllm-project/vllm · @lucas-santos-amd
  3. · Evidence links · GitHub search · Moonshot · pulls · vllm-project/vllm · @yiliu30
    Excerpt
    [Model] Load q/k/v scales for the quantized Kimi-K2.6 Register MLAAttention with the compressed-tensors KV-cache quantization method and map Kimi-K2.5/K2.6 checkpoint q/k/v scales into the model's
  4. · Evidence links · GitHub search · Moonshot · pulls · landfallbox/heimdall · @github-actions[bot]
    Excerpt
    - kimi: keeping retired models: kimi-k2.6, kimi-k2.7-code, kimi-k2.7-code-highspeed
  5. · Evidence links · GitHub search · Moonshot · pulls · datrics-ai/agent-market-docs · @kirikov
    Excerpt
    Judge: `crates/api/src/competition_judge.rs` — system account `COMPETITION_JUDGE`, default model `moonshotai/kimi-k2.6` (`COMPETITION_JUDGE_MODEL` overrides; one model per competition), the prompt in `build_prompt` (bands, untrusted data, steering lowers the
  6. · Evidence links · GitHub search · Moonshot · pulls · anomalyco/models.dev · @KevRojo
  7. · Evidence links · GitHub search · Moonshot · pulls · Mint-System/Odoo-Apps-Server-Tools · @LKooijman
    Excerpt
    model: moonshotai/Kimi-K2.6 input_tokens: 165441
  8. · Evidence links · GitHub search · Moonshot · commits · yetone/magpie · @yetone
  9. · Evidence links · GitHub search · Moonshot · commits · saikiranreddy18/toolnaut · @claude
  10. · Evidence links · GitHub search · Moonshot · pulls · anomalyco/models.dev · @models-dev-automation[bot]
  11. · Evidence links · GitHub search · Moonshot · pulls · anomalyco/models.dev · @models-dev-automation[bot]
    Excerpt
    -flash-lite, google/gemini-2.5-pro, meta-llama/llama-4-maverick, minimax/minimax-m2.5, minimax/minimax-m2.7, minimax/minimax-m3, moonshotai/kimi-k2.5, moonshotai/kimi-k2.6, moonshotai/kimi-k2.7-code, nvidia/nemotron-3.5-lightning, nvidia/nemotron-3.5-super-vl, nvidia/nemotron-3-nano-30b-a3b, openai/gpt-5.6, openai/gpt-6.1-sol-pro, poolside/laguna-s-2.1, poolside/laguna-xs-2.1, qwen/qwen3-235b-a22b-2507, qwen/qwen3.5-397b-a17b, qwen/qwen3-coder, ricursiveai/ripple-65b-preview, x-ai/grok-4
  12. · Evidence links · GitHub search · Moonshot · pulls · anomalyco/models.dev · @models-dev-automation[bot]
    Excerpt
    - deleted: `/home/runner/_work/models.dev/models.dev/providers/requesty/models/kimi-k2.6@eu.toml`
  13. · Catalogue listing · GitHub · MiniMax-AI/minimax-code · code
    Excerpt
    }; const presets = await parsePresetsForTest({ moonshotai: { name: 'Moonshot AI', npm: '@ai-sdk/openai-compatible', api: 'https://api.moonshot.ai/v1', models: { 'kimi-k3': k3, 'kimi-k2.6': { name: 'Kimi K2.6', tool_call: true, reasoning: true, reasoning_options: [{ type: 'toggle' }], }, 'kimi-k2.7-code': { name: 'Kimi K2.7 Code', tool_call: true, reasoning: true, reasoning_options: [], }, }, }, 'moonshotai-cn': { name: 'Moonshot AI China', npm: '@ai-sdk/openai-compatible', api: 'https://api.moonshot.cn/v1', models: { 'kimi-k3': k3 }, }, }); for (const providerId of ['moonshotai', 'moonshotai-cn'] as const) { const prese

    This code reference does not establish a working model identifier.

Signals 1

Specifications

Kimi K2.6 is a model from Moonshot with open weights, first recorded on 17 Sep 2026; its release date is not established. The context window is 262,144 tokens. Pricing is $0.65 per million input tokens and $3.41 per million output tokens.

Context window
262,144
Input
image · text · video
Output
text
Weights
Open
Input · USD / 1M tokens
$0.65
Output · USD / 1M tokens
$3.41
Cached input · USD / 1M tokens
$0.15

Questions

What is the context window of Kimi K2.6?

The context window is 262,144 tokens.

How much does Kimi K2.6 cost?

Pricing is $0.65 per million input tokens and $3.41 per million output tokens. Cached input is $0.15 per million.

Are the weights of Kimi K2.6 publicly available?

Yes. Moonshot publishes the weights.

When was Kimi K2.6 released?

Its release date is not established. Signal Forge first recorded it on 17 Sep 2026.