xAI launches Grok 4.6 with benchmark gains and lower-cost positioning for agent workloads

First recorded August 12, 2026 · Latest coverage August 12, 2026 · 1 source

xAI has released Grok 4.6, a new frontier model positioned for long-running agents, coding, and knowledge-work tasks. VentureBeat reports the model reached 61 on Artificial Analysis's Intelligence Index, surpassing Moonshot AI's Kimi K3 and tying OpenAI's GPT-5.6 Sol Max, while keeping API pricing at $2 per million input tokens and $6 per million output tokens.

Why it matters: The launch is a meaningful model-competition event because it combines reported benchmark improvement with aggressive price-performance positioning for enterprise agent use cases. If the reported gains hold in practice, Grok 4.6 could increase pressure on rivals to compete not only on top-end capability but also on the economics of running coding and autonomous-workload systems.

Sources

• SpaceXAI debuts Grok 4.6, overtaking Kimi K3's performance and matching GPT-5.6 Sol for world's third best on Artificial Analysis VentureBeat · August 12, 2026

Related stories

Independent events that offer a meaningful comparison, without implying that one caused the other.

• Google launches Gemini 3.7 Flash with coding and agent upgrades plus temporary lower API pricing
  August 13, 2026 · Independent comparison

  Google launched Gemini 3.7 Flash with coding and agent upgrades plus temporary introductory pricing of $0.75 per million input tokens and $3.75 per million output tokens, providing a direct pricing comparison for agent-oriented model economics.

• DeepSeek raises prices for V4 Flash and Pro as third-party tests show weaker real-world agent reliability
  August 16, 2026 · Independent comparison

  DeepSeek raised prices for V4 Flash and V4 Pro while third-party Composio testing showed V4 Flash completed 129 of 240 difficult agent-task runs, providing an independent comparison where weaker real-world reliability coincided with less aggressive pricing.

• OpenAI previews Ultrafast mode for GPT-5.6 Sol
  August 13, 2026 · Independent comparison

  As a separate comparison point, OpenAI's Ultrafast preview emphasizes up to 14-times faster GPT-5.6 Sol performance for lower-latency enterprise use, offering independent evidence that deployment efficiency is becoming a key competitive outcome.

Connections

Context and precedents—not claims of causation or corroboration.

• Anthropic launches Claude Sonnet 5 as a lower-cost model for agentic workloads
  June 30, 2026 · Contextual connection

  As a precedent, Anthropic launched Claude Sonnet 5 as a lower-cost model for agentic workloads, emphasizing stronger capabilities, lower pricing, and updated safety measures for enterprise automation use cases.

• xAI launches Grok Bot early beta for persistent workplace AI agents
  August 11, 2026 · Contextual connection

  As context, xAI had just launched Grok Bot in early beta as a persistent workplace agent product, a relevant product precedent for understanding Grok 4.6’s later positioning around long-running agent workloads.