SpaceXAI launches Grok 4.6 and 4.7 (batch) on OpenRouter
Esta notícia ainda só está disponível em inglês.
SpaceXAI added Grok 4.6 (batch) and Grok 4.7 (batch) to OpenRouter, priced at $1.6 per million input tokens and $4.8 per million output tokens.
Pricing and Context
Both models support up to 500,000 context tokens. The Grok 4.6 (batch) was released on 2026-08-12, while Grok 4.7 (batch) debuted on 2026-09-21.
Model Availability
The models are available immediately on OpenRouter under the x-ai provider. They are not open weights and do not have a free variant.
Technical Overview
Both Grok 4.6 (batch) and Grok 4.7 (batch) are built on the same architecture as earlier Grok models, featuring a transformer backbone with a 500,000 context window. They do not expose open weights and are only available through the x-ai provider on OpenRouter. The models are not listed on Hugging Face, as no hugging_face_id is present in the data.
Pontos principais
- SpaceXAI expands its Grok line with two new batch models.
- Both models carry identical pricing of $1.6 per million input and $4.8 per million output tokens.
- Each model supports 500,000 context tokens.