Kimi K3 API: Pricing, Benchmarks, and How to Access It (2026)

Kimi K3 is Moonshot AI's new flagship language model, released on July 16, 2026. At 2.8 trillion parameters it is the largest model ever announced with a planned public weights release, and the launch was loud enough to move stock markets in three countries. Behind the headlines sits a genuinely interesting model: a new attention architecture, a 1M-token context window, native image and video input, and benchmark results that put it next to the best closed models on several agentic tasks. This post covers what Moonshot published, what the numbers say, and where the claims still need independent confirmation.

TL;DR: Kimi K3 launched July 16, 2026 and is live on Unifically as moonshotai/kimi-k3 at $2.10 per 1M input tokens and $10.50 per 1M output, with full weights promised by July 27. Headline specs: 2.8T parameters (16 of 896 experts active per pass), 1M-token context, 131k default output, thinking always on. It posts 88.3% on Terminal Bench 2.1 and 91.2% on BrowseComp, and ranks #10 on Arena's Text board at 1486 Elo, 21 points behind Claude Fable 5 at #1. Thinking effort is locked to max at launch.

What Moonshot released

Kimi K3 is a from-scratch architecture, not a K2 fine-tune. The concrete specs from Moonshot's release:

• 2.8 trillion parameters. The first model in the 3-trillion class with a planned open release. Moonshot has held the open-model scale frontier for nine of the past twelve months, and K3 extends that run.
• Kimi Delta Attention (KDA) plus Attention Residuals. A hybrid linear attention design meant to keep information flowing through long sequences and deep layers. It is what makes the 1M-token window practical rather than decorative.
• Sparser Mixture of Experts. K3 activates 16 of 896 experts per forward pass under Moonshot's Stable LatentMoE framework. Moonshot reports roughly 2.5x the scaling efficiency of Kimi K2: more capability per unit of compute.
• 1M-token context. A single response defaults to 131k output tokens and can be raised to the full 1M. Context caching is automatic; no cache IDs or TTL settings.
• Native vision, including video. Images and video files go into the same request as text. The model uses screenshots as a feedback loop while coding, which shows up in its frontend and game-development results.
• Thinking always on. K3 runs in thinking mode on every request. The reasoning_effort parameter accepts only max at launch; low- and high-effort modes are promised in later updates.

The full model weights are scheduled for release by July 27, 2026, alongside a technical report. Until then, K3 is API-only.

The benchmarks

Moonshot published a comparison table against Claude Fable 5, GPT 5.6 Sol, Claude Opus 4.8, GPT 5.5, and GLM-5.2, all at max reasoning settings. Credit where due: they included benchmarks they lose.

Where K3 wins:

• SWE Marathon: 42.0%, the top score in the table, ahead of Claude Fable 5 at 35.0%. This is the long-run software engineering test, and it is K3's strongest claim: sustained multi-hour coding with minimal supervision.
• Terminal Bench 2.1: 88.3%, ahead of Fable 5 and Opus 4.8 (both 84.6%), just behind GPT 5.6 Sol at 88.8%.
• BrowseComp: 91.2% and DeepSearchQA: 95.0 f1, both first in the table. K3 is currently the strongest published model at agentic web search.
• OmniDocBench: 91.1% and SpreadsheetBench 2: 34.8%, both first. Document and structured-file work is a quiet strength.

Where it loses:

• HLE-Full: 43.5% against 53.3% for Fable 5. On the hardest frontier knowledge test, the gap is real.
• FrontierSWE: 81.2% against Fable 5's 86.6%.
• GDPval-AA v2: 1668 Elo against 1760 for Fable 5 and 1748 for GPT 5.6 Sol, on economically valuable knowledge work.

On Arena's crowd-voted Text board, K3 sits at #10 with 1486 Elo as of July 23, on 3,619 votes, wedged between Gemini 3 Pro at 1486 and GPT 5.6 Sol (xhigh) at 1485. Every model above it is closed. Claude Fable 5 holds #1 at 1507. The rating is still marked preliminary and has already moved several places since launch week, but for a model whose weights are days from being public, that neighborhood is unprecedented.

Independent numbers back the picture: Artificial Analysis measures 93.5% on GPQA Diamond and places K3 fourth of 187 models on its Intelligence Index at a score of 57.

What the open release means

If the weights land on July 27 as promised, K3 becomes the first open model at this scale, and the gap between open and closed frontier models shrinks to weeks rather than a year. That matters for three groups:

• Teams with data that cannot leave the building. A near-frontier model you can run in your own infrastructure changes what regulated industries can automate.
• Fine-tuners. A 2.8T MoE is expensive to serve, but an open flagship gives labs and enterprises a base no previous open release matched.
• Everyone negotiating API pricing. Open availability puts a ceiling on what closed providers can charge for comparable capability.

The caveat: 2.8 trillion parameters is not a model you run on a workstation. Even with 16-of-896 sparsity, self-hosting K3 is a multi-node project. For most teams the open weights matter less as a deployment path and more as bargaining power and insurance.

How to read the claims

Three things should stay open questions for now:

• The weights are a commitment, not a download. Moonshot says July 27. Until the files are on Hugging Face, K3 is a proprietary API model with an announcement attached.
• Most benchmark numbers are Moonshot's own. The table looks honest, and the Arena rank and Artificial Analysis results support it, but independent replication of the agentic results has not happened yet.
• Max-only thinking skews cost and latency. Every K3 request runs full reasoning today. Simple chat and extraction pay latency they do not need, and Artificial Analysis measures output speed at 62 tokens per second, below the median for its class. Judge production fit after the lower effort modes arrive.

Kimi K3 API pricing

K3 bills per token on Unifically, quoted per million:

Context caching is automatic, so a coding agent that resends a growing transcript every turn pays the $0.21 cache-read rate on most of its input, not the full $2.10. One number to anchor on: an agent step sending 50,000 input tokens and returning 5,000 costs about $0.16. Remember that reasoning tokens bill as output and K3 reasons at max on every request, so real output volume runs higher than the visible answer text. Live rates for every model are on the pricing page.

How to access Kimi K3 today

• Unifically. K3 is live as moonshotai/kimi-k3 on the OpenAI-compatible /v1/chat/completions endpoint, so existing OpenAI SDK code ports with a base-URL swap. A single response defaults to 131k output tokens and can be raised to the full 1M, streaming returns reasoning and answer text as separate deltas, and structured output uses strict json_schema. Tool calling, tool_choice: required, and dynamic tool loading all work through the standard request. On the platform K3 runs text in, text out today. You can also try it in the K3 playground first.
• Moonshot's own Kimi API. K3 is also live as model id kimi-k3 on Moonshot's OpenAI-compatible endpoint, with the same 131k default output and separate reasoning deltas.
• The Kimi apps. K3 runs in the Kimi assistant, Kimi Code, and the agent products, with max thinking effort by default.
• Self-hosting after July 27. Once the weights are published you can run K3 on your own hardware, but plan for a multi-node deployment; a 2.8T MoE does not fit a workstation.

A first call looks like this:

curl -X POST https://api.unifically.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "moonshotai/kimi-k3",
    "messages": [
      {"role": "user", "content": "Plan a migration from Express to Fastify for this service and list the risks in order."}
    ]
  }'

The same call in Python is a base-URL swap on the OpenAI SDK:

from openai import OpenAI

client = OpenAI(
    base_url="https://api.unifically.com/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="moonshotai/kimi-k3",
    messages=[{"role": "user", "content": "Plan a migration from Express to Fastify and list the risks in order."}],
)
print(response.choices[0].message.content)

Frequently asked questions

Does Kimi K3 have an API?

Yes. Kimi K3 is live on Unifically as moonshotai/kimi-k3 through the OpenAI-compatible /v1/chat/completions endpoint, and on Moonshot's own Kimi API as kimi-k3. Existing OpenAI SDK code works with a base-URL swap.

How much does the Kimi K3 API cost?

On Unifically, $2.10 per 1M input tokens, $10.50 per 1M output tokens, and $0.21 per 1M cached input tokens. Reasoning tokens bill as output, and K3 reasons at max effort on every request.

Is the Kimi K3 API free?

No. K3 bills per token with no subscription. A typical agent step of 50,000 input tokens and 5,000 output tokens costs about $0.16, so small experiments cost cents.

When do the Kimi K3 weights come out?

Moonshot has committed to publishing the full weights, with a technical report, by July 27, 2026. Until the files exist, K3 is API-only.

Which model should you use?

K3's launch profile is long coding runs, web research, and document work, and its 42.0 on SWE Marathon plus 91.2 on BrowseComp make it the pick when the run is long, research-heavy, or output-heavy — and it is the cheapest per token of the three flagships this month. If you want the strongest model on the depth tests instead, Claude Fable 5 leads the Arena Text board at #1, wins the knowledge-heavy benchmarks, and runs adaptive thinking rather than max-only. GPT 5.6 Sol edges K3 on Terminal Bench 2.1, runs far faster, and has an effort dial that goes down as well as up. The full head-to-heads: GPT 5.6 Sol vs Kimi K3 and Claude Fable 5 vs Kimi K3. All three run through the same endpoint, alongside every other language model on the platform.

We will update this post when the weights land, when independent benchmarks arrive, and when the lower thinking-effort modes change the latency story.