⚡Google shipped three Flash models.

Advertise here | 6-min Read

22 ChatGPT Agents Built for Every Marketing Job

Most marketers use ChatGPT to do general research and then call it an AI strategy. The ones outperforming them are deploying specialized agents built for specific jobs.

We put together 22 plug-and-play ChatGPT marketing agents that handle the work eating your week, each with built-in instructions and structured outputs ready to go in under 5 minutes.

Subscribe to Marketing Against the Grain and get all 22 free.

Inside you'll find:

• Competitive intelligence agent that visits competitor websites and builds detailed comparison matrices automatically

• Customer feedback analyzer that ranks improvement opportunities by business impact

• Social listening specialist that monitors brand mentions and flags reputation risks before they escalate

• Campaign optimization agents that handle attribution analysis and surface what is actually driving results

Your competitors are already running agents like these.

Get 22 ChatGPT Marketing Agents free when you subscribe to Marketing Against the Grain today.

Get The Guide

Claude Can Now Learn by Watching You

AI CHEAT SHEET

⚡Google shipped three Flash models. 3.5 Pro still is not one of them.

Google released three new Gemini models on July 21 — Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber — and used the same announcement to confirm that Gemini 3.5 Pro is still not shipping. The flagship has now missed multiple target dates. The three Flash models are real product work, not a distraction — Flash is where most enterprise AI volume actually runs.

The details:

• Gemini 3.6 Flash — the new default production tier. Upgraded reasoning and coding versus 3.5 Flash, same speed profile. Replaces 3.5 Flash as the recommended starting point for most enterprise workloads.

• Gemini 3.5 Flash-Lite — optimised for throughput and cost efficiency. The cheapest Gemini tier available, designed for high-volume classification, summarisation, and embedding-adjacent tasks where cost matters more than reasoning depth.

• Gemini 3.5 Flash Cyber — security-tuned, restricted to governments and trusted partners only. The first Google model explicitly scoped to defensive cybersecurity use. Released the same week JADEPUFFER and the Azure OpenAI 9.9 EoP patch dominated the security news cycle.

• Gemini 3.5 Pro update from the same announcement: still not shipping. Google did not give a new date. The stopgap Flash release is now confirmed to be exactly that — a stopgap.

Why it matters: The Flash tier is where the volume is. Most enterprise AI calls are routine tasks that do not require flagship reasoning — and Google's Flash pricing at $1.50/$9 per million tokens is the most aggressive mid-tier offer from any major Western lab. Teams that have been waiting for 3.5 Pro before committing to Google's stack should make a different decision: build on Flash now, migrate workloads when Pro ships and benchmarks hold up, and treat Pro as an upgrade rather than a prerequisite

• first organised industrial action in Asia directly attributed to AI automation in a manufacturing context.

Why it matters: Apple using Qwen in China is the most consequential AI governance story of the week — and it received far less attention than Gemini's third delay. It establishes a precedent: the world's most valuable company, with the most sophisticated AI safety and privacy positioning in consumer technology, is deploying a Chinese AI model to 1.4 billion users because the alternative was not operating in China at all. The regulatory divergence between China and the West is no longer a policy conversation. It is a product architecture decision.

Choose your AI model stack for Q3 — before the deadlines choose it for you

Prompt: You are an AI infrastructure architect. Three deadlines are converging in the next 40 days: DeepSeek legacy aliases die July 24, Kimi K3 full weights drop July 27, and Claude Sonnet 5 introductory pricing ends August 31. The model stack that made sense in June looks different today — and it will look different again in September.

Here is our current setup: [describe your workloads, which models you use, approximate monthly token volume, and current monthly spend].

Help me make three decisions before August 31:

1. The DeepSeek migration — if we use deepseek-chat or deepseek-reasoner in any pipeline, what exactly needs to change before July 24? Write the specific model ID updates, the test cases I should run before cutover, and the rollback plan if V4 does not behave the same way on our workloads. Do not assume V4 is a drop-in replacement — test it first.

2. The Kimi K3 evaluation — full weights drop July 27. If we want to evaluate it for any production workload before August 31, what is the minimum viable test? Five prompts that represent our hardest real use cases, not toy examples. For each: what does passing look like, and what failure would disqualify K3 from that workload entirely?

3. The Sonnet 5 August 31 decision — at $2/$10 until August 31, then $3/$15 from September 1: which of our workloads should we lock in on Sonnet 5 before the price change, which should we migrate to GPT-5.6 Terra at $2.50/$15, and which should we move to Gemini 3.6 Flash at $1.50/$9? Build the comparison table. Factor in the new tokenizer's 1.0–1.35x token overhead on Sonnet 5 — the nominal price change is 50%, but the effective cost change may be different.

End with a calendar: three specific actions, three specific dates, and the engineer responsible for each. No open items after August 31.

AI news highlights

• DeepSeek V4 lands Thursday — and legacy aliases die the same day — deepseek-chat and deepseek-reasoner return errors after the July 24 cutover. Migrate your model IDs today. This deadline does not move.
• Microsoft and Mistral signed a multibillion-dollar joint compute deal — Vera Rubin GPU-backed data centres in Europe — Mistral Medium 3.5 and Mistral OCR 4 added to Microsoft Foundry. Europe's AI infrastructure is being built with US GPU capacity and a French model stack.
• Kimi K3 full weights drop Friday — MIT license, 2.8T parameters, no restrictions — independent benchmarks expected over the weekend. If it competes with Fable 5 at open-weight pricing, the closed frontier premium argument weakens permanently.
• Defence AI attracted over $3 billion in July alone — Shield AI $1.5B Series G, Helsing €1.8B, Anduril-Archer partnership. Military autonomy is now competing with frontier model labs for investor capital.
• Anthropic spent $1.97M lobbying in Q2 2026 — outspending NVIDIA, nearly matching Oracle's $2M — OpenAI spent $1.2M. Combined AI lab lobbying hit $3.17M for the quarter, up 23% from Q1. The regulatory war has a budget.
• Fireworks AI raised $1.5B at $17.5B valuation — for inference infrastructure, not model development — the infrastructure layer beneath AI products is attracting frontier-lab-scale capital. Inference is now an asset class.
• ByteDance shipped FlowMimic — a mask-free video editor using temporal warped flow fields — no masks, no keyframes, just natural language edits applied to video. The AI video editing category is moving faster than the generation category.
• Claude Sonnet 5 introductory pricing: 40 days left at $2/$10 — standard $3/$15 from September 1. New tokenizer adds 1.0–1.35x overhead. Recalibrate before the switch, not after.

Trending AI tools

• • Kimi K3 — 2.8T parameters, MIT license, 1M token vision context — full weights Friday
• Wispr Flow — Dictate in any app — writes in your voice, auto-edits, 100+ languages
• OpenCode — Open-source, model-agnostic coding agent — air-gapped, no vendor dependency
• Retrace — Debug AI agents by replaying and forking runs at the exact point of failure
• Granola — AI meeting notes on top of any call tool — no bot, no permissions needed
• Framer Agents — Design, write, and organise your site with agents — Claude Code and Codex compatible
• Claude Cowork Mobile — Delegate and monitor long agent sessions from iOS and Android
• Osaurus — Open-source agents that run 100% locally on your Mac — no cloud, no data exposure.

That’s a Wrap

SPONSOR US

Get your business in front of over 90k+ AI professionals

8020AI is the world’s #1 AI Newsletter, Read by 90k+ professionals from leading companies such as Google, OpenAI, Meta, and Microsoft.

We've assisted in promoting Over 500 AI-Related Products. Will yours be the next?

What We Can Offer:

• Launch an Advertising Campaign

• Introduce New Product or Features

• Other Business Cooperation

Get in touch today.

Or Email our founder Alamin at [email protected]

FEEDBACK

How was your experience with 8020AI today?

Login or Subscribe to participate in polls.

If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.