G

Grok 4.6

SpaceXAI · Released Aug 2026

pending

SpaceXAI's next-generation 2-trillion-parameter model, completing training on July 21, 2026 and targeting an August 7 public launch. It scales up from Grok 4.5's 1.5T to ~2T parameters and aims to match or exceed Kimi K3's capability while preserving Grok's trademark token efficiency and ~80 TPS speed. Positioned as an Opus-class coding and agentic model trained alongside Cursor data, it will ship in Cursor, Grok Build, and the API at expected ~$2/$6 per 1M tokens. Pre-GA: no independent benchmarks exist yet; evaluation is provisional until the model ships and third-party testing validates the claims.

Is it right for you?

Good for

  • Expected to lead on token efficiency and per-task cost among frontier models, continuing Grok 4.5's ~4x token savings on coding tasks
  • Trained alongside Cursor IDE data for strong repository-scale coding and agentic task performance
  • Targets ~80 TPS throughput with custom C/C++ inference for fast iteration loops

Not good for

  • Strictly capped at 500K context, not 1M+ — Anthropic and Google offer larger windows for long-document work
  • EU availability may lag behind US launch, as was the case with Grok 4.5
  • No independent benchmarks or third-party evaluation available yet — all claims are vendor projections

Pricing

Input

$2.00 / 1M tokens

Output

$6.00 / 1M tokens

Context

500K tokens (expected, based on Grok 4.5)

View full pricing

No verdict changes yet

The clock starts day one — changes land here as our verdict evolves.

Verification log

  • Pricing— No changes

    Automated agent

How we evaluate