Moonshot AI Models on AtlasCloud | Kimi
Atlas Cloud hosts the full Kimi lineup via the MoonshotAI API, from K2-Thinking for deep reasoning to K2.6 for agentic coding. All pay-as-you-go, 262K context.
Moonshot AI is developed by Moonshot AI. Atlas Cloud (operated by Atlas Cloud AI LLC) provides access to it and does not own it. All trademarks belong to their respective owners.
Large Language Models by Moonshot AI
Power chat, reasoning, and agents at scale with leading large language models, served fast and affordably on Atlas Cloud.
Moonshot AI Models API Pricing Details
Compare standard vs. our pricing across every Moonshot AI model.
| Model | Standard Price (USD) | Our Price (USD) | Discount | |
|---|---|---|---|---|
| Kimi K3 | $3/$15per 1M tokens1048.6K context | $3/$15M in/outper 1M tokens1048.6K context | — | View |
| Kimi K2.7 Code | $0.95/$4per 1M tokens262.1K context | $0.95/$4M in/outper 1M tokens262.1K context | — | View |
| Kimi K2.6 | $0.95/$4per 1M tokens262.1K context | $0.95/$4M in/outper 1M tokens262.1K context | — | View |
| Kimi K2.5 | $0.6/$3per 1M tokens262.1K context | $0.49/$2.5M in/outper 1M tokens262.1K context | — | View |
Explore models from other providers
Instantly explore and experiment with 400+ production-ready models in the Atlas Playground. Start customizing with one click.
Moonshot AI API Use Cases You Can Build on Atlas Cloud
Kimi's agent swarm and long-horizon execution capabilities let teams run tasks that would take days of human effort in a single automated session. Teams use the M-series alongside K2-Thinking to cover everything from autonomous code changes to multi-document research at scale.
Legacy Codebase Modernization
Engineering teams use Kimi K2.6 to run long-horizon coding agents that autonomously overhaul production codebases over extended multi-hour sessions. In a documented example, K2.6 rewrote an 8-year-old financial matching engine over 13 hours and delivered a 185% throughput improvement without human intervention between commits. Atlas Cloud's pay-as-you-go pricing makes it practical to run these extended agentic sessions without capacity commitments.
Parallel Document Batch Processing
Operations teams use Kimi K2.6's 300-agent swarm to process large document batches in parallel. A single orchestration run matched one CV against 100 job roles and produced 100 fully customized resumes as output. The same pattern applies to contract review, compliance checks, and any workflow where a fixed input needs to be evaluated against a large, variable set of targets.
Deep Reasoning for Complex Analysis
Research and legal teams use Kimi K2-Thinking for multi-step analysis problems that require extended internal reasoning. The model supports up to 200 to 300 sequential tool calls per session, looping through reason-call-reason cycles without human prompting between steps. On Atlas Cloud it is priced at $0.6 per million input tokens and shares the 262K context window with the rest of the Kimi lineup.
Automated Research Paper Production
Academic and content teams use Kimi K2.6 to turn source documents into full research outputs. In a demonstrated run, K2.6 converted an astrophysics paper into a 40-page research paper, a structured dataset with over 20,000 entries, and 14 astronomy-grade charts in a single session. This reduces the turnaround on literature-to-output workflows from weeks to hours.
Business Prospecting at Scale
Growth and sales teams use Kimi K2.6 swarms to identify prospects and generate outreach assets in parallel. One example run identified 30 retail stores in a target city without websites and generated a landing page for each. The same pattern works for lead enrichment, competitive landscape mapping, and any task that combines discovery and content generation at list scale.
Visual Document and Code Analysis
Product and data teams use Kimi K2.5 and K2.6's native vision capabilities to process image and video inputs alongside text in the same API call. The MoonViT encoder handles diagrams, screenshots, UI mockups, and document scans without external preprocessing. This is useful for pipelines that convert visual specifications directly into code, or extract structured data from image-heavy documents.
Render your enterprise vision into reality with Atlas Cloud AI.
Contact SalesFrequently Asked Questions about Moonshot AI Models
Kimi K2.6 is MoonshotAI's latest open-source multimodal LLM, released in April 2026 under a Modified MIT license. It runs a Mixture-of-Experts architecture with 1 trillion total parameters and 32 billion active during inference. It is designed for agentic coding, long-horizon task execution, and multi-agent swarm orchestration.
Kimi K2.6 scales to 300 sub-agents executing up to 4,000 coordinated steps in a single run. Kimi K2.5 on Atlas Cloud supports swarm execution with up to 100 sub-agents. Tasks are dynamically decomposed into parallel, domain-specialized subtasks for fully autonomous output.
Kimi K2-Thinking uses deep chain-of-thought reasoning with up to 200 to 300 sequential tool calls per session. The model reasons, calls a tool, interprets the result, calls another tool, and continues this loop without human input. It is suited for multi-step logical inference, complex math, and problems where extended internal reasoning improves accuracy.
Yes. Kimi K2.5 and K2.6 include MoonViT, a 400-million-parameter vision encoder that processes images and video natively. You pass image or video inputs directly in the API call alongside text without external preprocessing. This supports visual analysis, document understanding, and image-to-code generation workflows.
Yes. Kimi K2.6 is released under a Modified MIT license, which permits commercial use. Open weights are available on HuggingFace for self-hosted deployments. Atlas Cloud also provides K2.6 via API for teams that prefer managed access without infrastructure overhead.
Kimi K2.6 scores 80.2% on SWE-Bench Verified and 54.0% on Humanity's Last Exam with tools, outperforming GPT-5.5 on both benchmarks. It also leads on BrowseComp at 83.2%, above GPT-5.4. These results come at roughly 80% lower cost per million tokens than GPT-5.5.
Kimi K2.5 is priced at $0.49 per million input tokens and $2.5 per million output tokens on Atlas Cloud. Kimi K2-Thinking and K2-Instruct-0905 run at $0.6 per million input tokens with the same output rate. Check the Atlas Cloud Kimi K2.6 model page for its current specific pricing.
Explore More Families
Kling 4.0
Developed by Kuaishou, Kling V4 is an video model built for professional production. It pairs stable complex motion with a 21:9 ultrawide format and targeted editing for expressions, camera angles, backgrounds, visual styles, and dialogue. Atlas Cloud is preparing Day-0 access through one OpenAI-compatible key with transparent pay-as-you-go pricing for flexible integration. Start building today.
Seedance 2.5
Seedance 2.5 API is now available on Atlas Cloud! It gives developers ByteDance's newest video model. It generates up to 30 seconds of native video in a single pass from text, a single image, or as many as 50 multimodal references, with synchronized audio and in-frame multilingual text. On Atlas Cloud you reach it through one key, with subject consistency and improved physics keeping long shots coherent. (Update: Seedance 2.5 1080P API Is Available NOW!)
Wan 3.0
Wan 3.0 API is the next generation of Alibaba's Wan video family, built to push long-form generation, multi-reference control, and audiovisual quality to new heights. Atlas Cloud already hosts Wan 2.7, 2.6, and 2.5, and Wan 3.0 runs on the same unified key with no separate setup. Start building today. Scroll down to the showcase to see what Wan 3.0 can create.
MiniMax H3
MiniMax H3 is MiniMax's multimodal video family for text, image, and reference guided creation. Across supported routes, it preserves subjects from reference media, offers flexible aspect ratios, and pairs generated sound with visuals through H3 Developer, with output profiles selected by endpoint. Atlas Cloud unifies the family behind one OpenAI-compatible key with transparent pay-as-you-go pricing from the standard rate of $0.038 per second. Start building today.
GPT Image 2.5
The gpt-image-2.5 family from OpenAI gives developers a choice of Flare and Sunburst for production image workflows. Render at arbitrary resolutions up to 3840x2160 and select from five quality tiers, including xhigh and max, to match specific output requirements. Atlas Cloud provides ready-to-use REST inference with no cold starts and standard pricing from $0.004 per generation. Start building today.
Seedream 5.0 Pro
Seedream 5.0 Pro API gives developers ByteDance's controllable image editing model on Atlas Cloud. It places edits precisely with anchors and coordinates, separates images into editable layers, fuses multiple references, and matches exact colors and materials, with multilingual text at 2K and 3K. On Atlas Cloud you reach it through one key!
Seedance 2.0
Seedance 2.0 is ByteDance’s production video model for precise shot creation. Turn prompts into video, animate a first-frame image with optional last-frame guidance, or shape results with reference media and optional web search. Atlas Cloud brings these workflows into one unified API with transparent pay-as-you-go pricing and one OpenAI-compatible key. Start building today.
Kling V3.0
From Kuaishou, the Kling V3 API family supports text, image, and reference video workflows. Transfer motion while preserving character identity, guide generation with visual references, and edit videos through natural language commands. Atlas Cloud offers Standard and Pro endpoints for different production needs. Already creating with Kling 3.0? Kling V4 is now available on Atlas Cloud. Explore it on the [Kling V4](https://www.atlascloud.ai/models/kling-v4) family page.
Grok Imagine Video
Grok Imagine Video gives developers access to xAI’s video family for creating, animating, extending, and editing clips. Generate 1 to 15 second videos from prompts, guide people, objects, or styles with up to seven reference images, and work at 480p or 720p. Atlas Cloud brings these capabilities together through OpenAI-compatible endpoints with transparent pay-as-you-go pricing. Start building today.
Vidu Q4 Preview
Vidu Q4 is Vidu's latest flagship video model, designed for image-to-video and reference-to-video generation. It can animate a single still image and create automatic scene transitions, giving developers two input paths for different product workflows. Atlas Cloud provides dedicated API endpoints with published parameter schemas at the verified standard price of $0.04 per second. Start building today.
Gemini Omni 1.1 Flash
Gemini Omni 1.1 Flash is Google DeepMind’s natively multimodal video family for flexible production workflows. Edit existing footage through text instructions, maintain visual consistency with up to 10 reference images and 3 video clips, or extend a shot in chainable segments to 40 seconds. Access every workflow on Atlas Cloud with one OpenAI-compatible API key, pay-as-you-go pricing, and Day-0 availability. Start building today.
Nano Banana 2.1
Nano Banana 2.1 is Google's image model family built on Gemini 3.6 Flash. It delivers Pro-level generation and editing at Flash-level speed, supports natural-language edits and merges across 1 to 14 input images, and maintains consistency for up to 4 characters and 10 objects. Through Atlas Cloud, developers can access the family with one OpenAI-compatible key, reliable uptime, and Day-0 availability. Start building today.
Recommended Articles
Guides, tutorials, and product updates to help you get the most out of Atlas Cloud.
