Tool· Updated August 1, 2026
AI pricing keyword trends
What people are actually typing into search about AI model pricing, coding agents, and hardware costs right now: 45 on-niche topics pulled from 397 raw autocomplete suggestions, sized by how many variants rolled into each one. This is a frequency signal from real search suggestions, not a verified ad-platform search-volume figure.
45
Topics mapped
From 397 raw suggestions
16
Model-pricing topics
The largest single category
GPT-5.6 pricing
Top topic by weight
Weight 29
20
Off-niche lines dropped
Noise excluded, not down-weighted
The keyword cloud
| Topic | Category | Autocomplete-frequency weight |
|---|---|---|
| GPT-5.6 pricing | Model pricing | 29 |
| Claude API cost | Model pricing | 26 |
| Claude Opus pricing | Model pricing | 22 |
| Gemini pricing & API | Model pricing | 19 |
| DeepSeek V4 | Model pricing | 18 |
| Nvidia GB300 | Hardware pricing | 17 |
| Claude Fable pricing | Model pricing | 15 |
| DDR5 memory prices | Hardware pricing | 14 |
| OpenAI pricing | Model pricing | 14 |
| Anthropic pricing | Model pricing | 13 |
| AI coding agents | Coding agents & tools | 12 |
| Claude Skills | Coding agents & tools | 12 |
| Codex CLI | Coding agents & tools | 11 |
| Cursor pricing | Coding agents & tools | 11 |
| GPU prices | Hardware pricing | 11 |
| CXMT | Hardware pricing | 10 |
| GitHub Copilot pricing | Coding agents & tools | 10 |
| Mistral Large pricing | Model pricing | 9 |
| Claude Code | Coding agents & tools | 8 |
| CoreWeave | Infrastructure | 8 |
| Grok pricing | Model pricing | 8 |
| LLM API cost | Model pricing | 8 |
| GPU cloud | Infrastructure | 7 |
| OpenRouter | Coding agents & tools | 7 |
| LLM gateway | Infrastructure | 6 |
| Agent harness | Coding agents & tools | 5 |
| AI data centers | Infrastructure | 5 |
| China AI model pricing | Model pricing | 5 |
| Kimi & Moonshot | Model pricing | 5 |
| Windsurf pricing | Coding agents & tools | 5 |
| B200 GPU price | Hardware pricing | 4 |
| vLLM | Coding agents & tools | 4 |
| LLM evaluation & benchmarks | Benchmarks & concepts | 3 |
| Qwen models | Model pricing | 3 |
| Autonomous AI agents | Benchmarks & concepts | 2 |
| LLM observability | Infrastructure | 2 |
| AI model legal status | Industry & policy | 1 |
| AI pricing models | Model pricing | 1 |
| AI researchers leaving labs | Industry & policy | 1 |
| AI visibility tools | Benchmarks & concepts | 1 |
| AI/agentic model law | Industry & policy | 1 |
| Generative vs agentic vs predictive AI | Benchmarks & concepts | 1 |
| Mira Murati / Thinking Machines | Industry & policy | 1 |
| Subscription vs API pricing | Model pricing | 1 |
| What is an AI model | Benchmarks & concepts | 1 |
Topics by category
The 45 topics split into 6 categories. Model pricing dominates, reflecting how much of current AI-related search is people trying to work out what a model actually costs to use, whether by API or subscription. For the actual verified prices behind these topics, see the model release tracker and subscription-vs-API calculator.
| Category | What it covers | Topics |
|---|---|---|
| Model pricing | Per-token and subscription pricing for frontier and open-weight models. | GPT-5.6 pricing, Claude API cost, Claude Opus pricing, Gemini pricing & API, DeepSeek V4, Claude Fable pricing, OpenAI pricing, Anthropic pricing, Mistral Large pricing, Grok pricing, LLM API cost, China AI model pricing, Kimi & Moonshot, Qwen models, AI pricing models, Subscription vs API pricing |
| Coding agents & tools | AI coding agents, CLIs, harnesses, and the pricing plans behind them. | AI coding agents, Claude Skills, Codex CLI, Cursor pricing, GitHub Copilot pricing, Claude Code, OpenRouter, Agent harness, Windsurf pricing, vLLM |
| Hardware pricing | GPU, memory, and accelerator prices that set the floor for inference cost. | Nvidia GB300, DDR5 memory prices, GPU prices, CXMT, B200 GPU price |
| Infrastructure | GPU cloud, data centers, gateways, and the providers running the stack. | CoreWeave, GPU cloud, LLM gateway, AI data centers, LLM observability |
| Benchmarks & concepts | Evaluation, visibility tooling, and foundational AI-model questions. | LLM evaluation & benchmarks, Autonomous AI agents, AI visibility tools, Generative vs agentic vs predictive AI, What is an AI model |
| Industry & policy | Personnel moves, research labs, and the legal status of AI models. | AI model legal status, AI researchers leaving labs, AI/agentic model law, Mira Murati / Thinking Machines |
For the coding-agent and hardware topics mapped here, see the Claude Skills and agent directory, cost-per-task ranking, inference providers directory, and DDR5 price tracker.
Get each breakdown before it makes the rounds
You get one email when a new source-backed analysis goes live: what AI agents actually cost, which models are worth running, and what the benchmarks really mean. No hype.
No spam. Unsubscribe anytime.
Frequently asked questions
- What is this keyword cloud based on?
- It is built from 397 raw autocomplete suggestions scraped across seed terms in AI model pricing, coding agents, and hardware economics, grouped into 45 on-niche topics and sized by how many raw variants rolled into each one. It is a frequency signal from real search suggestions, not a verified Google Ads search-volume number.
- Why were some keywords dropped?
- 20 raw lines were dropped because they were off-niche noise picked up incidentally by the autocomplete scrape (stock tickers, unrelated names, generic filler) rather than genuine AI pricing or model-economics demand. Only lines that matched a named on-niche topic were kept.
- What is the single biggest topic right now?
- GPT-5.6 pricing, with a weight of 29, meaning that many distinct raw autocomplete variants (different phrasing, different intents) rolled up into it.
- How often is this updated?
- The underlying source list is refreshed periodically from the keyword-radar autocomplete scrape and rebuilt through the same grouping and filtering script, so the cloud reflects a point-in-time snapshot dated to when it was last regenerated.