xAI Releases Grok 4.6: Top-Tier Performance at Unbeatable Cost

xAI has officially released the new generation model Grok 4.6. The model ranks among the top tier of large language models in multiple benchmark tests and achieves significant performance improvements while maintaining the same pricing as its predecessor, making extreme cost-effectiveness its core competitive advantage. Both Elon Musk and Nvidia have highly praised its performance.

已确认

• 基准表现：Grok 4.6 took first place in the GDPVal-AA v2 benchmark with an ELO score of 1753, surpassing Fable 5 Max and GPT-5.6 Sol Max. In the Artificial Analysis Intelligence Index, it scored 61, tying with GPT-5.6 Sol and just behind Claude Opus 5 and Claude Fable 5. It also ranked first in workload tests with a $5 budget.
• 定价策略：The model's pricing remains consistent with the previous generation, Grok 4.5, at $2 per million input tokens, $6 for output tokens, and $0.50 for cached input. Analyst Gavin SB Baker noted that its performance rivals Fable 5 Max, but its overall cost is 85% lower (80% cheaper for input, 88% cheaper for output), making it even cheaper than Anthropic's Sonnet 5.
• 平台接入：The model is now available on Grok Build, Cursor, Grok Bot, and the xAI API. Official announcements stated that Cursor and Grok Build users will receive double token allowances within the first week of release.
• 硬件支持：Nvidia officially confirmed that Grok 4.6 was trained and runs on the GB300 NVL72 system.

为什么重要

In the current fiercely competitive landscape of large language models, Grok 4.6 has successfully achieved the "Pareto optimality" hinted at by Elon Musk. It not only demonstrates cutting-edge intelligence in complex scenarios like real-world tasks and agentic knowledge work (e.g., the AA-Briefcase test) but also drastically lowers the barrier to using top-tier AI models through aggressive pricing. This cost-effectiveness-focused strategy directly disrupts the market positioning of competitors like OpenAI and Anthropic, offering developers and enterprise users a more cost-efficient SOTA option.

2026-08-12 ~ 2026-08-13 · 77 related posts

Full story(4 episodes)→

• Episode 1: Grok 4.6 Spotted in Cursor Then Pulled, Unconfirmed by xAI(2026-08-11, 7 posts)
• Episode 2: Grok 4.6 Released with Multi-Tool Coding Benchmark(2026-08-11, 4 posts)
• Episode 3: xAI Releases Grok 4.6: Top-Tier Performance at Unbeatable Cost(2026-08-12, 77 posts)
• Episode 4: Developers Test Grok 4.6: Fast Programming Speed and High Cost-Performance(2026-08-13, 2 posts)

• xAI Releases Grok 4.6 Model API — legit_api · 2026-08-12
• xAI Releases Grok 4.6: Focuses on Long-Running Agents, Matches GPT-5.6 — scaling01 · 2026-08-12
• xAI Launches Grok 4.6: 1.5T Parameters with Major Coding & Agent Upgrades — Daniel_Farinax · 2026-08-12
• User Trolls with Fake Grok 4.6 Benchmarks, Begs Elon Musk for Student AGI Plan — felpix_ · 2026-08-12
• Grok 4.6 available in Cursor and API with 2x usage for the first week — Daniel_Farinax · 2026-08-12
• Grok 4.6 Released? AI Community Trolls With Fake Model Updates — kimmonismus · 2026-08-12
• Grok 4.6 Matches GPT-5.6 Sol on Intelligence Index, Leads in Coding Benchmarks — kimmonismus · 2026-08-12
• [source] Grok 4.6 ties GPT-5.6 on intelligence index with top-tier agentic performance at lower cost — ArtificialAnlys · 2026-08-12
• Grok 4.6 Debuts Strong on AA-Briefcase, Trailing Only Claude Opus 5 — ArtificialAnlys · 2026-08-12
• Grok 4.6 Released, Beats GPT-5.6 on GDPVal-AA v2 Benchmark — XFreeze · 2026-08-12
• Introducing Grok 4.6: Significant Frontier Intelligence Leap at Same Price — andersonbcdefg · 2026-08-12
• [source] Grok 4.6 Takes #1 Spot on GDPVal-AA Benchmark with 1753 Elo — elonmusk · 2026-08-12
• xAI Releases Grok 4.6: Combines Opus-Class Intelligence with Low Cost — soleio · 2026-08-12
• Elon Musk Praises Grok 4.6: Ranks #1 Across Multiple Benchmarks — elonmusk · 2026-08-12
• xAI Officially Launches Grok 4.6: Major Performance Boost at the Same Price — xiaosun86 · 2026-08-12
• [source] Grok 4.6 Crushes Competitors on Value; Grok 4.7 to Integrate SpaceX Data — GavinSBaker · 2026-08-12
• xAI Releases Grok 4.6, Now Available on Cursor and Grok Build — mark_k · 2026-08-12
• Grok 4.6 Tops Multiple Benchmarks While Maintaining Previous Generation Pricing — XFreeze · 2026-08-12
• Grok 4.6 Test: On Par with Sol at One-Fifth the Cost — daniel_mac8 · 2026-08-12
• Grok 4.6 is Now Available in Cursor — prasenx · 2026-08-12
• Grok 4.6 Benchmarks Leaked: Major Performance Jump at Same Price — mark_k · 2026-08-13
• Grok 4.6 Hits the API at $2 per Million Input Tokens — XFreeze · 2026-08-13
• Grok 4.6 Drops and Hits LiveBench Soon, Priced Same as 4.5 — bindureddy · 2026-08-13
• Grok 4.6 Stuns in Eval: Matches Rivals at a Fraction of the Cost — daniel_mac8 · 2026-08-13
• Grok 4.6 Matches GPT-5.6 Sol in AI Index at Less Than Half the Cost — Angaisb_ · 2026-08-13
• Grok 4.6 Lands in Cursor and Grok Build with 2x Usage for the First Week — XFreeze · 2026-08-13
• Grok 4.6 Matches Competitor Performance at 85% Discount — GavinSBaker · 2026-08-13
• Grok Offers 85% Discount Over OpenAI with Similar Performance — GavinSBaker · 2026-08-13
• Three Frontier Models in One Day: Grok, Qwen, and DeepSeek — Yuchenj_UW · 2026-08-13
• Grok 4.6 Tops Artificial Analysis Agentic Index, Tying Claude Opus 5 Max — XFreeze · 2026-08-13
• Grok 4.6 Matches Competitor Performance at 85% Discount; Next-Gen to Include Cursor Data — altryne · 2026-08-13
• AI Community Jokes About 'Crazy Wednesday' with Fake Grok 4.6 and DeepSeek V4 Pro — vista8 · 2026-08-13
• Grok and Cursor show Pareto dominance as Anthropic deemed 'unreliable' partner — GavinSBaker · 2026-08-13
• xAI Launches Flagship Grok 4.6, Rivaling GPT-5.6 and Claude 5 at Lower Cost — 智东西 · 2026-08-13
• Grok 4.6 Matches GPT 5.6 at 32% Lower Cost, Open Models Overpriced — downingARK · 2026-08-13
• Grok 4.6 Scores 61 on Intelligence Index, Joining the Frontier — elonmusk · 2026-08-13
• Grok 4.6 Integrated into Devin, Cursor, and Other Dev Tools — elonmusk · 2026-08-13
• NVIDIA Congratulates xAI on Grok 4.6: Trained on GB300 NVL72 — elonmusk · 2026-08-13
• Joke on xAI Release Cadence: Grok 4.6 Arrives, Fable 5.1 Nowhere in Sight — patience_cave · 2026-08-13
• Mercor Reports Grok 4.6 as One of the Most Cost-Efficient Frontier Models Tested on APEX — GavinSBaker · 2026-08-13
• Elon Musk Pushes Grok: Major Leap in Real-World Task Performance — elonmusk · 2026-08-13
• Grok 4.6 Released: Better Performance at Half the Price of Rivals — NinaDSchick · 2026-08-13
• Grok 4.6 Nearly Matches Claude Fable 5 on Agentic Benchmark at a Fraction of the Cost — ArtificialAnlys · 2026-08-13
• Artificial Analysis Details AA-Briefcase: A Private Benchmark for Long-Horizon Agentic Workflows — ArtificialAnlys · 2026-08-13
• Grok 4.6 Matches Claude Fable 5's Intelligence at 5-8x Lower Token Cost — rohanpaul_ai · 2026-08-13
• Grok 4.6 tops AI models under $5 workload budget, showing high intelligence per dollar — XFreeze · 2026-08-13
• Grok 4.6 tested on bug bench: outperforms predecessor, becomes new default — PawelHuryn · 2026-08-13
• xAI Launches Grok 4.6: Major Performance Boost at Same Price, Now in Hermes Agent — NousResearch · 2026-08-13
• xAI Launches Grok 4.6 Across Cursor, API with 2x Token Promo — aman_madaan · 2026-08-13

1 near-duplicate retellings: kimmonismus