· 2 min read
The Gradient — 2026-09-24
- Company Blog
OpenAI’s GPT-6 Astra is now powering Harvey’s legal drafting, delivering structured, context‑aware documents. This shift lets attorneys focus on strategy rather than boilerplate.
Harvey turns legal context into stronger drafts with GPT-6 Astra
What happened: Harvey integrated the GPT-6 Astra model into its drafting workflow, producing legal documents that automatically incorporate relevant case context and maintain a consistent structure.
Why it matters: By offloading routine drafting to an AI that understands legal nuances, lawyers can allocate more time to strategic analysis, risk assessment, and client counseling, raising the overall quality of legal services.
Key stats: • Drafting time cut by roughly 40% on average.
• Document error rates dropped about 30% compared with manual drafts.
• User satisfaction surveys show an 85% approval rating among participating attorneys.
Source: https://openai.com/index/harvey-from-context-to-confidence-with-astra
How invideo improves color grading 3x with GPT‑6 Astra
What happened: InVideo integrated the new GPT‑6 Astra model, enabling the platform to plan edits with far‑greater precision, triple the speed of color correction and grading, and generate 50 bespoke visual effects in a single day.
Why it matters: Faster, higher‑quality color grading reduces post‑production time and costs, giving creators more creative flexibility and helping studios meet tight deadlines.
Key stats: • Color grading speed up 3× • 50 custom effects produced per day • GPT‑6 Astra powers edit‑planning precision
Source: https://openai.com/index/invideo-builds-with-gpt-6-astra
Ringg’s AI agents resolve up to 65% of customer calls with OpenAI
What happened: Ringg deployed GPT‑5.6‑based agents that operate via voice, chat, WhatsApp, and web, achieving resolution of up to 65% of inbound customer calls.
Why it matters: Automating such a high share of calls reduces staffing pressure, shortens wait times, and improves customer satisfaction while slashing operational expenses.
Key stats: • 65% of calls resolved without human agents
• Multilingual support across voice, chat, WhatsApp, web
• 90% lower cost compared to using GPT‑4.1
Source: https://openai.com/index/ringg
Airbnb widens access to GPT-6 Astra and OpenAI frontier models
What happened: Airbnb announced that its engineering organization can now use GPT‑6 Astra alongside OpenAI’s frontier models for internal projects. The rollout includes API access, sandbox environments, and dedicated support.
Why it matters: By leveraging the most advanced language models, Airbnb expects faster bug resolution, smarter system architecture, and shorter time‑to‑market for new features. The initiative also signals a broader trend of enterprises adopting next‑gen AI to boost productivity.
Key stats:
- Expected 30% reduction in average bug‑fix time.
- Up to 2× faster prototype iteration for new product ideas.
- Pilot program includes 150+ engineers across three global hubs.
Better prompt caching for GPT-6
What happened: GPT‑6 introduces an upgraded prompt‑caching layer that raises cache‑hit rates across typical workloads.
Why it matters: Higher hit rates cut inference latency and token‑usage costs, while new diagnostics and explicit breakpoints give developers finer control over cache behavior.
Key stats: • Cache‑hit rate up by 30% vs GPT‑5 • Latency reduced by ~15% • Cost per 1 M tokens down ~12%
Source: https://openai.com/index/better-prompt-caching-for-gpt-6