Grok 4.6 Aces ARC-AGI-1, Matches GPT-5.6 at a Fraction of the Cost
elonmusk · x · 2026-08-14
According to verified ARC-AGI data, xAI's Grok 4.6 scored 87.5% on ARC-AGI-1 at just $0.30/task. On the hardest tier, ARC-AGI-3, it performed comparably to GPT-5.6 Sol but cost $5.6K versus Sol's $15.2K.
More from Models
- GLM-5.2 Gains Attributed Entirely to Post-Training, Base Model Unchanged — ivan_bezdomny · 2026-08-14
- Developer Builds AI Dungeon Master with DeepSeek: 1000+ Turns for Under $2 — zacurryy · 2026-08-14
- Opinion: Post-Training Is All You Need for LLM Advancements — IridiumEagle · 2026-08-14
- DeepSeek Launches V4-Pro: 1.4T Parameters with Major Agent Upgrades — drdanielbender · 2026-08-14
- Google Exec Confirms Gemini 4 Pre-training, 3.5 Pro Likely Skipped — haider1 · 2026-08-14
- Grok 4.6 Ranks #1 on CursorBench for Real-World Coding — kevinnbass · 2026-08-14