Grok 4.7 hits Amazon Bedrock: 500K context, 2x output tokens for the gains
AWS ML Blog · rss · 2026-09-29
AWS ML Blog details xAI's Grok 4.7 on Amazon Bedrock: 500K token context, four reasoning-effort levels, and Responses/Chat Completions/Converse APIs. Trained with longer RL on harder tasks, it emphasizes self-verification for long-horizon agents. Artificial Analysis scores it Intelligence Index 46 (vs 44), Coding Agent 56 (vs 47), hallucination rate down to 29% — but output tokens roughly doubled (81K/task). Ships with a new safeguard stack (strongest refusals/jailbreak resistance xAI has tested). On Bedrock: cross-region inference profiles, OpenAI SDK compatibility, implicit prompt caching, Guardrails, structured outputs, and default/priority/flex service tiers.
More from Infra
- Mobile AI Factory: agents auto-build on-device SLMs with 19x less GPU time — pmttyji · 2026-11-16
- Stratechery's Fictional Obituary: Imagining Jensen Huang's 2026 Exit From Nvidia — Gregory_C_Allen · 2026-11-16
- vLLM serves DeepSeek V4.1-Flash at 954K tokens/s per GPU, $38.96/GPU-hour on GB300 NVL72 — bookwormengr · 2026-11-16
- Colossus 2's 1.2M B200s could generate ~$400B/year in tokens at DeepSeek Flash rates — bookwormengr · 2026-11-16
- BIS issues rule blocking remote access to US AI chips via third-country data centers — teortaxesTex · 2026-11-16
- Qdrant 1.20.1: 50% fewer cold-storage reads, ~2x faster quantization, agent-ready setup — qdrant_engine · 2026-11-16