Thinking Machines Inkling Small
The smaller Inkling (276B total, 12B active parameters), close to the larger model on many benchmarks at lower cost and latency, reading text, images and audio.
From Thinking Machines, Inkling Small is a cloud text model. It is multimodal: alongside a text prompt it accepts images and audio, then replies with generated text. Its context window handles up to 524K input tokens and up to 262K output tokens, and it supports adjustable reasoning effort for harder problems. Generation can be tuned with system instructions, temperature, and top-p sampling. It runs through OpenRouter using your own API key, from $0.45 per million input tokens.
- Released
- Jul 2026
- Pricing
- $0.45 / 1M in · $1.20 / 1M out
- Inputs
- Text, Images, Audio
- Context window
- 524K in · 262K out
- Reasoning
- Adjustable effort
- Controls
- System prompt, Temperature, Top-p
Samples
Generated with Inkling Small via OpenRouter. 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 letting one program ask another for data or actions, like fetching weather from a service. It defines how to ask, what to say, and what answer returns, so different apps communicate without knowing each other's inner workings.
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 authentication provider immediately. With only two people, your scarcest resource is time, and user authentication is a deep, risky rabbit hole of password security, session management, multi-factor authentication, breach recovery, and compliance obligations that will consume weeks of engineering and expose you to catastrophic liability if done poorly. A hosted service lets you ship today, stay focused on your actual product, and inherit enterprise-grade security without becoming experts in cryptography. The single strongest argument against this is that building in-house eliminates recurring subscription costs, removes vendor lock-in and the risk of a third-party breach or outage affecting your users, and gives you complete customization and data sovereignty over credentials as you scale.
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 budgets defy traditional forecasting because the underlying economics behave more like weather than manufacturing. First, compute and data costs scale unpredictably; a pilot running smoothly on small models can suddenly require tenfold infrastructure investment once production traffic hits, and cloud pricing for specialized GPUs shifts with supply constraints. Second, talent and vendor markets move fast; the consultants who architected last year’s system now command premium rates, while open-source tools adopted can suddenly need expensive commercial licenses to meet security standards. Third, technical obsolescence is relentless; a model family that seemed sufficient six months ago may be superseded by a more capable version, forcing unplanned retraining and integration rather than simple maintenance. To regain control, treat your AI spending as a portfolio rather than a project. Establish monthly governance reviews comparing actual consumption against usage metrics, not just invoices. Build flexible budgets with tiered commitments—commit to fixed costs for core platforms but leave headroom for experimental workloads. Demand transparent pricing from vendors and avoid long-term contracts for rapidly evolving services. Most importantly, tie every dollar to a clear business outcome and sunset initiatives that do not show measurable return within a defined horizon, so unpredictability is contained by strict discipline rather than avoided altogether.
Thinking Machines
Thinking Machines Lab builds open-weight multimodal models, including Inkling and Inkling Small, which read text, images and audio, and Tinker, a fine-tuning API.
thinkingmachines.ai ↗