Grok 4.7

	Grok 4.7(x.ai)
	598 points by meetpateltech 22 days ago \| 517 comments
	tl;dr: xAI released Grok 4.7, a larger base model trained with extended RL on longer, harder tasks, priced at $2/$6 per million input/output tokens (same as 4.6). It posts competitive scores across coding (CursorBench 46.3%, Terminal-Bench 38%), knowledge work, and legal/clinical benchmarks, though it trails Fable 5.1 on several. xAI also touts a new safeguard stack with strong jailbreak resistance and dual-use safety, including a 62.4% score on LatchBio's biosafety benchmark.
	HN Discussion: ↓Skeptical that xAI is happy with 4.7 results given price/margin and delays; competitors will surpass it ~Shares hands-on testing results showing incremental improvements in specific tasks like image-to-HTML ↓Grok 4.7 is too slow, expensive, and burns tokens to chase benchmarks without clear real-world gains ↑Optimistic about xAI's release cadence and expects bigger improvements with Grok 5 ↓Competitors like Astra and DeepSeek offer better value or performance, making Grok less compelling