Model Overview & Capabilities
Grok 4.6 (high) is a frontier artificial intelligence model engineered by SpaceXAI, officially verified in June 2026 and evaluated on the Artificial Analysis Intelligence Index v4.3 with an overall score of 44.0. Architected as a proprietary API service, the system operates across a context window of 500k tokens, allowing enterprise developer agents and autonomous code evaluation tools to ingest comprehensive repository structures, cross-language modules, and multi-layered documentation hierarchies without degradation of semantic coherence.
Under standardized software engineering benchmarks—including SWE-bench Verified, Terminal-Bench 2.0, Aider Polyglot, and LiveCodeBench—Grok 4.6 (high) exhibits rigorous multi-step problem solving, deterministic SEARCH/REPLACE diff compliance, and robust tool-use navigation in sandboxed execution environments. Its reasoning capabilities are calibrated to minimize hallucination rates while handling complex syntax transformations, boundary conditions, and continuous integration workflows.
Economically and operationally, Grok 4.6 (high) generates output tokens at a median throughput of 60 tokens per second, with an observed initial latency of 44250 milliseconds. Artificial Analysis rates its estimated cost per task at $1.86 USD, backed by an API token pricing structure of $2.0 per million input tokens and $10.0 per million output tokens (resulting in an effective blended rate of $4.0 per million tokens).