NeoSmith Basic

NeoSmith Basic is positioned as a specialized reasoning model designed for complex logic and chain-of-thought tasks, making it a natural fit for building advanced reasoning agents. The model accepts both text and image inputs while producing text outputs, and it ships with native tool calling support, adjustable temperature control, and attachment handling. Its closed-weight design is paired with a generous context window and substantial output capacity, allowing long-running agent loops and multi-step reasoning workflows without frequent truncation. The capability set is intentionally lean: it enables reasoning and tool use but does not expose a dedicated structured output mode, so teams that need strict JSON responses will need to validate outputs themselves.

In practical terms, NeoSmith Basic offers a cost-effective price point relative to its long context and reasoning feature set, and prompt caching is supported with cheap read rates while cache writes are free, which helps when reusing large system prompts or retrieved context across many calls. The model is served through the NeoSmith router endpoint, integrates with the standard OpenAI-compatible SDK, and is targeted at developers constructing agentic systems rather than one-off chat completions. Teams that need a reasoning-capable model with tool use, image understanding, and a very large working context will find the fit appropriate, while those requiring first-class structured output or open weights for self-hosting should look elsewhere.