Kimi K3 pricing in 2026
What one task costs on Kimi K3
Per-million-token prices are hard to feel, so here are typical token counts for common jobs priced with Kimi K3. Figures are USD at the rates above and exclude prompt caching and batch discounts.
| Task | Tokens (in / out) | Cost |
|---|---|---|
| Code review of a 500-line pull request | 60,000 / 4,000 | $0.240 |
| Debug one failing test | 40,000 / 6,000 | $0.210 |
| Generate a landing page in one shot | 3,000 / 30,000 | $0.459 |
| Draft a 1,200-word article | 2,000 / 2,500 | $0.043 |
| Answer a question from a 20-page document | 25,000 / 800 | $0.087 |
| Multi-step agent run (10 tool calls) | 250,000 / 25,000 | $1.13 |
| All of the above, once | $2.16 |
Cheaper at the same or better capability: GLM-5.3 FlashX runs the same set of tasks for $0.226 instead of $2.16.
Kimi K3 by Moonshot AI costs $3 per 1M input tokens and $15 per 1M output tokens — about 83.3k tokens per $1. Prices last verified 2026-09-19.
The verdict
Moonshot's reasoning flagship with an index of 44 at $3 / $15. Strong quality that undercuts OpenAI and Anthropic, though its $15 output price is pricier than Grok 4.6 or Gemini 3.8 Flash at similar quality.
A concrete example: sending 100,000 input tokens and getting back 20,000 output tokens with Kimi K3 costs about $0.60 at current prices.
Who is Kimi K3 for?
Kimi K3 is Moonshot's reasoning flagship — an index of 44 at $3 / $15 per 1M. It delivers top-tier reasoning and outstanding design work while undercutting OpenAI and Anthropic, and it's the best open-weight model for vibe coding and game dev.
Its main drawback is speed: Moonshot lacks the compute to serve it quickly through every route (direct access is faster). If you can wait for quality, K3 is excellent value; if you need speed, Grok 4.6 or Gemini 3.7 Flash respond faster.
Strengths & weaknesses
- Strength: Top-tier reasoning from Moonshot
- Best for: Design, Game-dev, Open-weight
- Weakness: Slow — Moonshot lacks the compute to serve it quickly, though direct access is faster.
Kimi K3 key facts
| Provider | Kimi pricing |
|---|---|
| Input price | $3 / 1M tokens |
| Output price | $15 / 1M tokens |
| Tokens / $1 | 83.3k |
| Intelligence / $1 | 36,667 |
| Context window | 1.0M tokens |
| Tier | frontier |
| Speed | Slow |
| Intelligence index | 44 / 100 |
| Released | 2026-07-16 |
| Last verified | 2026-09-19 |
What Kimi K3 delivers for your budget
For a coding workload, here is roughly what each monthly budget buys:
| Budget | Tokens / month | Months of workload |
|---|---|---|
| $20 | 3.5M | 5.3x |
| $50 | 8.7M | 13.3x |
| $100 | 17.3M | 26.7x |
How Kimi K3 compares to its closest rivals
Rather than repeating the full catalog on every page, here are the three current models nearest to Kimi K3 on the intelligence index — and why you might still pick one instead:
- Grok 4.6 (Grok) — $2 input / $6 output per 1M tokens. Why you might pick it: Frontier index at mid price.
- Step 5 Preview (Step) — $1 input / $2.70 output per 1M tokens. Why you might pick it: Near-frontier index for a fraction of the flagship price.
- Qwen3.8 Max (Qwen) — $2 input / $6 output per 1M tokens. Why you might pick it: Multimodal frontier capability from Alibaba.
Kimi K3 — frequently asked questions
How much does Kimi K3 cost?
Kimi K3 by Moonshot AI costs $3 per 1M input tokens and $15 per 1M output tokens.
How many tokens per dollar does Kimi K3 give you?
On a typical 1:3 input-to-output mix, Kimi K3 delivers around 83.3k tokens per dollar — an intelligence-per-dollar score of 36667.
What is Kimi K3's context window?
Kimi K3 supports a context window of 1.0M tokens.
Is Kimi K3 worth it in 2026?
Kimi K3 is Kimi's frontier tier model. Use the AI Intelligence Calculator to see how its value compares to every other model at your budget.