Grok 4.6 Matches GPT-5.6 Sol

Grok 4.6’s headline score matters, but its real story is cost-efficient execution across long-running agent workflows.

• –Scores 1,753 Elo on GDPval-AA v2 and 88.4% on Terminal-Bench v2.1
• –Matches frontier models across knowledge work, customer service, and terminal-based coding
• –Costs $2/$6 per million input/output tokens, versus $5/$30 for GPT-5.6 Sol
• –Completes AA-Briefcase tasks in roughly half the turns and one-quarter the input tokens of Claude Opus 5 Max
• –Benchmark leadership still needs validation in production workloads, where reliability, latency, safety, and tool integrations matter as much as composite scores