GPT-6 Astra API pricing is $10 per million input tokens, but a prompt over 272K tokens reprices the whole request to $20. Here is the real cost, and how it compares.
Sun Sep 06 2026 · 6 min read · 2 views
The advice you have read about MCP context cost is out of date. The old rule was simple. Every tool you connect gets injected up front, so a big connector spends a chunk of the window before you type
Sat Aug 22 2026 · 6 min read · 3 views
Prompt caching is sold as a saving. It can also be the single largest line on an inference invoice, and the failure is quiet, because a cache write succeeds whether or not anything ever reads it back.
Fri Aug 21 2026 · 6 min read · 6 views
Set temperature=0.7 on Claude Opus 5 and the request fails. The API returns a 400 error, not a warning and not a silently ignored parameter. The same happens with topp, topk, and with the manual
Fri Aug 21 2026 · 6 min read · 22 views
Agents degrade as context grows, well before the window is full. Here is the token budget that fixes it, and what each fix costs you in latency.
Thu Aug 20 2026 · 5 min read · 1 views
Anthropic and EPFL showed AI agents passing self-spreading instructions through shared prompt files. The fix is one paragraph, and it works today.
Thu Aug 20 2026 · 5 min read · 1 views
Feature tables do not predict cost. Rank AI agent frameworks in 2026 by churn: what each one broke, when, and whether your code survived it.
Thu Aug 20 2026 · 5 min read · 3 views
Most teams can see their agent failing in production. Fewer than half can catch it beforehand.
Wed Aug 19 2026 · 6 min read · 2 views
Most advice on agent memory treats it as a storage question. Pick a vector store, summarize when the window fills, move on. That misses the bill.
Wed Aug 19 2026 · 6 min read · 6 views
Your agent calls the same tool repeatedly and never returns an answer. The advice you will find first is to set maxiterations, which caps your bill without addressing the underlying defect.
Wed Aug 19 2026 · 6 min read · 3 views
AutoGen is not getting new features. The project README states it plainly: "AutoGen is now in maintenance mode. It will not receive new features or enhancements and is community managed going
Wed Aug 19 2026 · 5 min read · 2 views
OpenAI's Assistants API stops working on August 26, 2026. That is six days from today. After that date, calls to /v1/assistants, /v1/threads and /v1/threads/runs return an error, according to OpenAI's
Wed Aug 19 2026 · 5 min read · 2 views
Anthropic cancelled the September 1 Claude Sonnet 5 price increase and made $2/$10 permanent. Here is what that changes for your model routing and budget.
Wed Aug 19 2026 · 5 min read · 7 views
GitHub Copilot's promotional AI credits stop on August 31, 2026. Nothing about how your team writes code changes on September 1. The invoice does.
Tue Aug 18 2026 · 7 min read · 7 views
Stripe has finalized a deal to acquire OpenRouter for more than $7 billion, Bloomberg reported on August 16, 2026. If your product calls a large language model through OpenRouter, nothing breaks this
Tue Aug 18 2026 · 7 min read · 1 views
OpenAI now sells a model built to write exploits, and almost nobody can buy it. GPT-5.6-Cyber sits inside a tier called Daybreak Red, behind identity checks, legal paperwork and, from September 1,
Sun Aug 16 2026 · 5 min read · 4 views
There is now a public database of more than 1,600 court decisions worldwide in which someone filed AI-fabricated material and a judge responded. It is the largest documented record of AI being used
Sun Aug 16 2026 · 5 min read · 3 views
Browsers and chat apps are now happy to hand an AI agent your payment details and let it check out on your behalf. Nobody has told you what happens when the agent buys the wrong running shoes, buys
Sun Aug 16 2026 · 6 min read · 0 views
Two AI labelling laws switched on together on August 2, 2026 — Article 50 of the EU AI Act and California's AI Transparency Act — and almost every explainer written about them is aimed at compliance
Sun Aug 16 2026 · 6 min read · 3 views
DeepSeek moved its entire API over to peak and off-peak billing at 16:00 UTC on August 16, 2026. Output generated by the flagship V4-Pro model now costs $3.96 per million tokens at peak, and $1.98 the
Sat Aug 15 2026 · 6 min read · 30 views