How does Grok 4.6 perform compared to other frontier models?

Grok 4.6 scored 61 points on the Artificial Analysis Intelligence Index, tying OpenAI's GPT-5.6 Sol. Only Anthropic's Claude Opus 5 (63 points) and Claude Fable 5 (62 points) rank higher. On the GDPval-AA v2 benchmark, Grok 4.6 ranks second with an Elo score of 1,753, completing complex tasks in about 53 steps versus roughly 103 for Claude Opus 5.

What is the pricing for Grok 4.6?

Grok 4.6 costs $2/$6 per million tokens, which is more than 60 percent cheaper than Claude Opus 5 ($5/$25) and GPT-5.6 Sol ($5/$30).

Where can I access Grok 4.6?

Grok 4.6 is available now through the API, Cursor, Grok Build, and partners like OpenRouter, Vercel, and Cloudflare. For the first week, x.ai is offering double the usage quota in Grok Build and Cursor.

Large Language ModelsAI Business & IndustryTHE DECODER4h agoAug 13, 2026, 04:01 JST3 min read

Grok 4.6 matches OpenAI's best model at 60% lower price

Send on LINE

Key takeaway

SpaceXAI's Grok 4.6 has matched OpenAI's top model on the Artificial Analysis Intelligence Index, scoring 61 points and ranking second on real-world knowledge-work benchmarks.
The model costs more than 60 percent less than competing frontier models from Anthropic and OpenAI, making it available immediately through multiple platforms and API partners.

3 Key Points

What happened
SpaceXAI's Grok 4.6 scored 61 points on the Artificial Analysis Intelligence Index, tying OpenAI's GPT-5.6 Sol and trailing only Anthropic's Claude Opus 5 (63 points) and Claude Fable 5 (62 points). On the GDPval-AA v2 benchmark, Grok 4.6 ranks second with an Elo score of 1,753, completing complex tasks in about 53 steps versus roughly 103 for Claude Opus 5.
Why it matters
Grok 4.6 is priced at $2/$6 per million tokens—more than 60 percent cheaper than Claude Opus 5 ($5/$25) and GPT-5.6 Sol ($5/$30)—making frontier-tier performance accessible at significantly lower cost. This narrows the pricing gap between SpaceXAI and established leaders in AI capability.
What to watch
Grok 4.6 is available now through the API, Cursor, Grok Build, and partners like OpenRouter, Vercel, and Cloudflare. For the first week, x.ai is offering double the usage quota in Grok Build and Cursor.

Summaries like this, in your inbox every morning.

Get news like this every morning — free

In Depth

Read the full story

SpaceXAI has released Grok 4.6, a new frontier model that achieves performance parity with OpenAI's most capable public offering while undercutting it significantly on price. According to the Artificial Analysis Intelligence Index—which aggregates multiple benchmarks into a single score—Grok 4.6 achieved 61 points, tying OpenAI's GPT-5.6 Sol. Only two models rank higher: Anthropic's Claude Opus 5 at 63 points and Claude Fable 5 at 62 points. This represents a five-point gain over Grok 4.5, its predecessor.

The model demonstrates particular strength on agentic tasks, where AI systems complete multi-step workflows independently. On the GDPval-AA v2 benchmark, which measures real-world knowledge work on a computer, Grok 4.6 ranks second with an Elo score of 1,753, trailing only Claude Opus 5. The performance gap is substantial in terms of efficiency: Grok 4.6 completes complex tasks in roughly 53 steps, while Claude Opus 5 requires approximately 103 steps.

Pricing remains unchanged at $2/$6 per million tokens (input and output tokens respectively). This places Grok 4.6 more than 60 percent below Claude Opus 5, which costs $5/$25 per million tokens, and GPT-5.6 Sol at $5/$30 per million tokens—making frontier capability available at a fraction of competitors' rates.

Grok 4.6 is now accessible through the API, Cursor (a coding editor integrating AI), Grok Build (SpaceXAI's platform), and external partners including OpenRouter, Vercel, and Cloudflare. To drive initial adoption, x.ai is doubling usage quotas in Grok Build and Cursor for the model's first week of availability.

Context & Analysis

SpaceXAI's Grok 4.6 represents a significant competitive move in the frontier AI market. The model's 61-point score on the Artificial Analysis Intelligence Index ties OpenAI's GPT-5.6 Sol, positioning it among the world's most capable language models. What distinguishes Grok 4.6 is not just its performance tier but its pricing: at $2/$6 per million tokens, it undercuts Claude Opus 5 by 60 percent or more. This pricing strategy suggests SpaceXAI is pursuing market share in a segment dominated by Anthropic and OpenAI, where performance has historically commanded premium prices. The model's particular strength on agentic tasks—where AI systems independently manage multi-step workflows—aligns with an emerging category of work where businesses increasingly deploy AI. Grok 4.6's immediate availability through multiple platforms (API, Cursor, Grok Build, plus OpenRouter, Vercel, and Cloudflare) signals a deliberate push for rapid adoption, reinforced by the one-week promotion doubling usage quotas.

FAQ

How does Grok 4.6 perform compared to other frontier models?: Grok 4.6 scored 61 points on the Artificial Analysis Intelligence Index, tying OpenAI's GPT-5.6 Sol. Only Anthropic's Claude Opus 5 (63 points) and Claude Fable 5 (62 points) rank higher. On the GDPval-AA v2 benchmark, Grok 4.6 ranks second with an Elo score of 1,753, completing complex tasks in about 53 steps versus roughly 103 for Claude Opus 5.
What is the pricing for Grok 4.6?: Grok 4.6 costs $2/$6 per million tokens, which is more than 60 percent cheaper than Claude Opus 5 ($5/$25) and GPT-5.6 Sol ($5/$30).
Where can I access Grok 4.6?: Grok 4.6 is available now through the API, Cursor, Grok Build, and partners like OpenRouter, Vercel, and Cloudflare. For the first week, x.ai is offering double the usage quota in Grok Build and Cursor.

THE DECODERRead Original Article