Articles tagged "GPT-5.6"

August 2026: OpenAI's Ultrafast preview runs GPT-5.6 Sol on Cerebras at up to 750 output tokens/sec, 14x standard. Pricing undisclosed, access limited.

LLM 0.32, released Aug 4 2026, is llm CLI's biggest update yet: reasoning traces, server-side tools, a new SQLite log schema, and a Python messages API.

On July 30, 2026 OpenAI cut GPT-5.6 API prices: Luna drops 80% to $0.20 in / $1.20 out per 1M tokens, Terra drops 20% to $2 / $12, Sol is unchanged and gains Fast mode (up to 2.5x speed at 2x price). Full old-vs-new tables, monthly cost recalculations, and what to change in your routing design.

GPT-5.6 ships in three tiers — Sol $5/$30, Terra $2.50/$15, Luna $1/$6 (input/output per 1M). xAI Grok 4.5 is $2/$6 with a 500K context and an OpenAI-compatible API. A practical comparison plus per-use-case routing, prompt caching, and reasoning-effort cost tips for SMBs. Updated July 2026.

**OpenAI unveiled [ChatGPT Work](https://openai.com/chatgpt-work/) at a 2026-07-09 livestream** ([Axios](https://www.axios.com/2026/07/09/ai-openai-gpt-release) / [Bloomberg](https://www.bloomberg.com/news/articles/2026-07-09/openai-unveils-chatgpt-work-agent-to-field-tasks-for-hours)). It's a **GPT-5.6-powered workplace AI agent** that **gathers context across Slack / Microsoft Teams / Google Drive / SharePoint / email / calendars / CRMs** and produces documents, spreadsheets, presentations, and other work artifacts autonomously. The agent **breaks projects into smaller steps and runs for hours**, while users watch progress, answer questions, redirect it, and approve important actions. Users can **pull context from a specific app by typing `@`** followed by the app name (e.g. `@Slack`, `@GoogleDrive`). **Rollout**: web + mobile begins with Pro / Enterprise / Edu on 2026-07-09, expanding to Plus / Business within days. **The updated desktop app** (Mac / Windows) also ships **Chat / Work / Codex to every plan including Free**. GPT-5.6, released the same day ([CNBC](https://www.cnbc.com/2026/07/08/openai-expanding-gpt-5point6-ai-model-release-ending-government-limits.html) / [The Hill](https://thehill.com/policy/technology/5958647-openai-releases-gpt56-trump/)), is the substrate model for ChatGPT Work. Position: a **head-on competitor to [Claude Cowork's Web/Mobile expansion](../columns/claude-cowork-web-mobile-launch-2026-07)**, matching Cowork's HITL-notification + multi-app-integration approach with OpenAI's workplace-agent strategy. Joins [Devin](../columns/claude-fable-5-return-devin-cursor-2026-07) (developer-focused), [Hermes Agent](../columns/nous-portal-cloud-hermes-agent-2026-07) (vendor-neutral multi-model), and [Herdr](../columns/herdr-terminal-agent-multiplexer-2026-07) (terminal parallel ops) as the **late-2026 agent big four**.