Category
artificial-intelligence
-
AI agent unit economics in 2026: a cost model that survives audit
Agentic models burn 5 to 30 times the tokens of a chatbot, and Gartner expects cost per agentic workflow to rise more than fivefold through 2028. The human comparator, meanwhile, is unsourceable.
-
Claude computer use went GA on 19 August 2026: seven changes that break an existing agent loop
Anthropic made computer use generally available on 19 August 2026 and shipped a browser use toolset alongside it. The migration changes the request shape, flips zoom on by default, stops downscaling oversized
-
Claude Files API went GA on 19 August 2026 and the beta header you no longer need now silently pins you to the old
Anthropic made the Files API generally available on 19 August 2026. Requests that still send the files-api-2025-04-14 beta header keep the previous response format, a fork documented in exactly one place and in neither
-
Groq retired llama-3.3-70b-versatile on 16 August 2026 and points production teams at a Preview model
Groq's deprecation page names qwen/qwen3.6-27b as a replacement for llama-3.3-70b-versatile. The same page says Preview models should not be used in production. Here is what the swap actually costs.
-
Mistral shipped Agentic Search on 20 August 2026, and Libraries cannot run on the 1.1x EU endpoint
Mistral's Agentic Search launch claims a jump from 26.7% to 86% on FinanceBench. The 86% figure belongs to a Z.ai model, and the managed path that delivers it cannot run on Mistral's EU regional endpoint.
-
26 August 2026: OpenAI removes the Assistants API, and there is no automated thread migration
OpenAI's Assistants API is removed on 26 August 2026. The migration guide claims no lost feature support, then says there will be no automated tool for moving threads to conversations.
-
1.25x cache writes start billing on 21 August 2026: OpenAI's new dashboard, Azure's missing rate
OpenAI released a Prompt Caching dashboard on 20 August 2026 that reports cache reads per write. On GPT-5.6 that ratio decides whether caching cuts your input bill by 78 percent or raises it by 25 percent, and Azure
-
Zed 1.16.1 changed the default OpenAI model on 19 August 2026 while its price table lagged OpenAI by 22 days
Zed's 1.16.1 release on 19 August 2026 switched the default model for OpenAI subscription users to GPT-5.6 Sol. Its hosted price table still carries OpenAI's pre-30-July rates for Terra and Luna, so the published Luna
-
OpenAI shuts Agent Builder, Evals and reusable prompts on 30 November 2026
Three OpenAI developer surfaces die on 30 November 2026 after a single 40-word changelog entry. The export path for evals closes a month earlier, on 31 October, and OpenAI's own docs still disagree about where Agent
-
Vercel AI Gateway's 18 September 2026 audio cliff: TTS bills per UTF-8 byte, not per character
Vercel's 19 August 2026 changelog quotes Fish Audio text-to-speech at $15.00 per million characters. Fish Audio's own rate card says per million UTF-8 bytes. For Hindi or Chinese the two are not the same number.
-
Azure Document Intelligence v2.0 retires 31 August 2026: three Microsoft pages, three different migration targets
Microsoft's retirement notice for Azure AI Document Intelligence v2.0 points at v3.1. The Learn documentation updated 11 June 2026 does not mention v2.0 at all and points at v4.0. Both pages are live. Here is how
-
Chatbot Development Company: 2026 Costs to Ship a Production Support Bot in India
Outcome pricing at $0.99 a resolution costs 64 times the underlying model tokens. The published 2026 numbers, the resolution-rate denominator games, the Indic tokenisation tax, and the DPDP clause that should change