all the models — AI benchmark observatory
← Models

Grok 4.5

xAI · proprietary · grok-4.5

Grok 4.5 is SpaceXAI's frontier model for coding, agentic tasks, and knowledge work. Trained alongside Cursor across tens of thousands of NVIDIA GB300 GPUs, it uses large-scale reinforcement learning focused on multi-step software engineering and other technical work. The model supports text and image inputs, reasoning, function calling, structured outputs, and a 500K-token context window, and is served at 80 tokens per second.

AA-OmniscienGPQATerminal-Ben

Benchmark scores

BenchmarkScore
AA-Omniscience Index126.00
GPQA0.93
Terminal-Bench 2.10.83

Pricing

  • xAI$2.00 / $6.00

Input / output per 1M tokens

AA metrics

No Artificial Analysis link yet.

Arena Elo

  • vision1279
  • code1555
  • text1469