xAI: Grok 4.20 Multi-Agent

Grok 4.20 Multi-Agent is a paid model from xAI that accepts text, image, and file inputs and supports reasoning tasks. Its context window reaches 2 million tokens, which places it among the larger-context options currently available. It does not support tool use, and structured output support is unconfirmed, so workflows that depend on function calling or guaranteed schema-formatted responses will need a different model. At $1.25 per million input tokens and $2.50 per million output tokens, the pricing is competitive for a reasoning-capable model with this context capacity. The tradeoff is transparency: there is no independent benchmark coverage yet, so performance claims rest entirely on xAI's own representations. Buyers who need a large context window for document-heavy or multi-source reasoning tasks may find it worth evaluating, but those who require third-party performance validation before committing should wait for external benchmark results to emerge.