Inkling vs. Antares

Inkling is an open-weights multimodal AI model from Thinking Machines built to support customization, agentic workflows, coding, reasoning, vision, audio, and enterprise AI use cases. The model is a Mixture-of-Experts transformer with 975 billion total parameters, 41 billion active parameters, 256 routed experts per MoE layer, and six routed experts active per token. It supports context windows up to 1 million tokens and was pretrained on 45 trillion tokens across text, images, audio, and video. Inkling is designed as a broad foundation model rather than a narrowly optimized benchmark model, giving it balanced capabilities across reasoning, coding, factuality, instruction following, vision, audio, tool use, and safety. Its controllable thinking effort lets developers adjust how much computation and generated reasoning the model uses, helping teams balance quality, latency, and cost for different production needs. The model can run agentic coding tasks, use tools, create web apps, generate polished multi-page artifacts, reason over long contexts, and work through iterative refinement loops. For multimodal tasks, Inkling can process images, answer questions about visual content, transcribe and reason over audio, follow spoken instructions, and combine visual reasoning with code-based tools such as Python. Thinking Machines trained Inkling for calibration, instruction following, factual reliability, refusal behavior, and safety across multiple modalities, including evaluations for dangerous capabilities and human-AI threat vectors. Inkling is available on Tinker for fine-tuning, with 64K and 256K context options, an Inkling Playground for testing, cookbook recipes, and support for multimodal post-training workflows. Its full weights are available on Hugging Face, and deployment support is available through APIs and infrastructure partners such as TogetherAI, Fireworks, Modal, Databricks, Baseten, SGLang, vLLM, llama.cpp, and transformers.

Antares is a collection of open-weight security small language models crafted to detect vulnerabilities within large codebases. Featuring models such as Antares-350M and Antares-1B, these tools can be deployed locally or on-site, ensuring that proprietary source code remains secure while also reducing both inference expenses and runtime. The procedure starts with an outline of the vulnerability, which may include an advisory or a CWE category; from there, the model embarks on a detailed investigation similar to that of a human analyst, methodically looking for relevant code patterns, scrutinizing possible files, integrating new data, and adjusting its strategy when certain paths appear unproductive. This method allows the model to concentrate its resources on the files most likely to contain the identified flaws. In the end, Antares produces a prioritized list of source files that may be vulnerable, accompanied by a comprehensive trail of the exploration process that led to these conclusions, thereby simplifying the review and prioritization for teams. Furthermore, this functionality not only accelerates the vulnerability assessment process but also significantly strengthens the overall security framework of the development environment, fostering a culture of proactive security measures. Ultimately, organizations can benefit from improved efficiency and effectiveness in managing their code vulnerabilities.