Kimi K2.6
Moonshot's frontier open-weight model (1T parameters) built for long-horizon agentic coding, with a 262K context and native vision.
Kimi K2.6 is a cloud text model built by Moonshot AI. It is multimodal: alongside a text prompt it accepts images, then replies with generated text. Its context window handles up to 262K input tokens and up to 33K output tokens, and it supports adjustable reasoning effort for harder problems. Generation can be tuned with temperature and top-p sampling. It runs through Replicate using your own API key, from $0.95 per million input tokens.
- Pricing
- $0.95 / 1M in · $4.00 / 1M out
- Inputs
- Text, Images
- Context window
- 262K in · 33K out
- Reasoning
- Adjustable effort
- Controls
- Temperature, Top-p
Examples
Generated with Kimi K2.6 via Replicate. The same three prompts run against every text model in the catalog, shown verbatim — so the only thing that changes between two models’ answers is the model.
Explain what an API is to someone who has never written code. Do not use a restaurant or waiter analogy. Plain prose only — no headings, bullet points, or markdown formatting. Maximum 50 words.
An API is a set of rules that lets one program talk to another. It works like a translator, taking your request and turning it into instructions the other system understands, then bringing back the result you wanted, without you needing to know how that other system works inside.
A two-person startup is deciding whether to build their own user authentication or use a hosted provider. Give them a clear recommendation, then the single strongest argument against your own recommendation. Plain prose only — no headings, bullet points, or markdown formatting. Maximum 150 words.
Use a hosted provider like Auth0, Clerk, or Firebase Authentication. Building your own system requires deep security expertise, ongoing maintenance, and significant development time that a two-person startup cannot spare. Hosted solutions handle the critical complexities of password hashing, token management, breach monitoring, and compliance requirements out of the box, letting your tiny team focus on core product features that actually generate revenue and attract customers. The strongest argument against this recommendation is vendor lock-in and long-term cost. As your user base scales, hosted authentication pricing can grow substantially, and migrating away later becomes technically painful and risky. If the provider changes pricing, suffers extended downtime, or shuts down entirely, you may face an expensive emergency rebuild of a system you never fully understood, potentially disrupting users and damaging trust at the worst possible moment in your company's growth.
Write a briefing for a non-technical executive explaining why their company's spending on AI is hard to predict, and what they can do about it. Cover at least three distinct causes. Plain prose only — no headings, bullet points, or markdown formatting. Write exactly 205 words.
Artificial intelligence spending is inherently difficult to forecast for several interconnected reasons. First, the technology itself evolves unpredictably. A model that seemed state-of-the-art six months ago may become obsolete overnight, forcing unplanned investments in newer architectures or triggering costly retraining cycles that were not budgeted. Second, usage patterns defy traditional software economics. Unlike conventional enterprise software with fixed per-seat licenses, many AI services charge based on tokens processed, API calls made, or compute consumed. One viral product feature or unexpected surge in customer adoption can multiply cloud bills exponentially within days, making month-to-month costs swing wildly. Third, talent markets are extraordinarily volatile. The small pool of researchers and engineers who can genuinely advance AI capabilities commands compensation packages that fluctuate with competitive bidding wars, and retention bonuses or counteroffers can emerge suddenly, disrupting salary projections entirely. To manage this uncertainty, executives should implement rolling forecasts updated quarterly rather than annual budgets, negotiate spending caps and reserved capacity with cloud providers, and build modular architectures that allow swapping components without wholesale system rebuilds. Cross-training existing technical staff reduces dependency on scarce specialists. Most critically, leadership must treat AI investments as a portfolio of experiments with explicit kill criteria, ensuring that unpredictable costs attach to measurable learning rather than open-ended commitments.
Moonshot AI
Moonshot AI builds the Kimi family — frontier open-weight models with very long context, native visual understanding, and strong agentic coding.
www.moonshot.ai ↗