Requesty
Requesty Blog

AI Models

Every post tagged AI Models.

Looking for the longer pieces? Read our research on the coding agent economy or browse the data notes.

All173Best Practices81Integrations59Routing42Agents38Requesty Features33AI Models23Observability16Cost Optimization15Security13Pricing8Industry7Enterprise6AI Gateway6Infrastructure6

2026

23 posts

Jev, week two: 1,316 posts, four gateways giving it away, and an open weight clone…

Grok 4.7 and MiMo-V2.6 Pro launched four hours apart with the same benchmark score: one…

Open weight models are 55% of our tokens and 15% of our spend: what the 78.4% Vercel…

TypeSafe Jev explained: how it works, LLM differences and API pricing

Which of the smartest models can you run inside the EU, and what is the residency premium?

Claude in the EU: 47 region deployments, one quota wall, and how to fail over without…

DeepSeek V4.1 Flash retires V4 Pro by redirect: on 14 September your pinned model id…

GLM-5.3 Flash vs DeepSeek V4 Flash 0731: which flash model fits your API workload?

GPT-6 Astra is the best model you may not be allowed to call: gated access, the EU gap…

The list price is now a range: peak hours, promo windows and host floors

GPT-6 Astra scores 61 on the independent index, the same as Sol, at 2.5x the price

Record on Terminal-Bench, catastrophic on a private eval: Fable 5.1 and the case for your…

Fable 5.1 cut cache reads 75% to $0.25 per million: the launch line nobody screenshotted

Five open weight releases in nine days: GLM-5.3-Flash, Qwen3.8-Flash, Hy4 and the…

A 320B model trained on Chinese AI chips at 1/100 frontier price: the inference supply…

20 trillion tokens in 6 days: what the Ox Alpha stealth launch taught us about model IDs

Claude Opus 5 vs GPT-5.6 Sol: API speed, coding and cost

Free model IDs now ship with expiry dates: how to use promotional inference without…

36x the price for 22% more quality: the Pareto data that should decide your model mix

The supply side is exploding: 136 providers, 502 models, and 2,084 client apps in a…

No model stays #1: the leader's share of gateway traffic fell from 29% to 8% in eight…

Inside Sakana Fugu Ultra: We Reverse Engineered Its Multi Agent Architecture

Best AI Coding Model (2026): Benchmarks, Cost, and Real World Performance

Start building with Requesty

One line of code. 600+ models. Full control.

Speak to founders