Track AI model prices, token costs, inference trends, market heat and performance signals across the global AI economy.
Models are changing. Prices are moving. Compute costs are shifting. HotON.ai turns fragmented AI signals into structured market intelligence.
From model prices and usage trends to availability, latency and market heat — the whole AI economy on one screen.
AI pricing is no longer simple. HotON.ai helps builders, enterprises and investors understand how model costs move across providers and categories.
| Model | Region | Input | Output | Ctx |
|---|---|---|---|---|
| MINIMAX-M2.7 MiniMax | CN | $0.30 | $1.20 | 1000K |
| DEEPSEEK Novita AI | US | $0.14 | $0.28 | 1049K |
| QWEN3-CODER-FL Alibaba Cloud · Qwen | CN | $0.20 | $0.98 | 1000K |
| GEMINI-2.0-FLA Google | US | $0.08 | $0.30 | 1049K |
| GEMINI-2.5-FLA Google | US | $0.10 | $0.40 | 1049K |
| QWEN-PLUS-2025 Alibaba Cloud · Qwen | CN | $0.26 | $0.78 | 1000K |
| GPT-4.1-MINI-2 OpenAI | US | $0.40 | $1.60 | 1048K |
| GEMINI-3.1-FLA Google | US | $0.25 | $1.50 | 1049K |
| GROK-4-FAST xAI | US | $0.20 | $0.50 | 2000K |
| LLAMA-4-MAVERI Meta | US | $0.15 | $0.60 | 1049K |
| MINIMAX-01 MiniMax | CN | $0.20 | $1.10 | 1000K |
| MINIMAX-M2 MiniMax | CN | $0.26 | $1.00 | 1000K |
| GEMINI-2.5-FLA Google | US | $0.10 | $0.40 | 1049K |
| MINIMAX-M2.1 MiniMax | CN | $0.29 | $0.95 | 1000K |
Structured benchmarks for AI model prices, efficiency, inference cost and market momentum. News gets copied — indexes don't.
The cheapest model is not always the most efficient. HotON.ai compares models by total task cost, success rate, speed, stability and output quality.
| # | Model | Cost / Task | Context | Arena Elo | Efficiency |
|---|---|---|---|---|---|
| 01 | MINIMAX-M2.7 MiniMax | $0.001 | 1000K | — | 96 |
| 02 | DEEPSEEK Novita AI | $0.000 | 1049K | — | 96 |
| 03 | QWEN3-CODER-FL Alibaba Cloud · Qwen | $0.001 | 1000K | — | 96 |
| 04 | GEMINI-2.0-FLA Google | $0.000 | 1049K | — | 96 |
| 05 | GEMINI-2.5-FLA Google | $0.000 | 1049K | — | 96 |
| 06 | QWEN-PLUS-2025 Alibaba Cloud · Qwen | $0.001 | 1000K | — | 96 |
| 07 | GPT-4.1-MINI-2 OpenAI | $0.002 | 1048K | — | 96 |
Find the model that actually delivers the best result for the cost. See the full price-vs-intelligence frontier →
AI costs are shaped by more than model prices. Region, compute supply, energy cost, latency and availability all matter.
Illustrative map of provider regions — not a live feed.
Compare inference cost patterns across global regions.
Understand where AI infrastructure capacity is becoming more attractive.
Track how energy conditions may influence compute and inference pricing.
Discover when certain regions may become more cost-efficient for AI workloads.
HotON.ai helps the market understand the geography of AI cost.
Model launches, pricing changes, infrastructure shifts, policy updates, funding events and market movements — filtered from the noise.
Perplexity trusts GPT-6 Astra with end-to-end systems
OpenAI agents attacked RubyGems back in May
How to Build an AI Software Factory: Agents That Open, Review, and Merge PRs
OpenAI agents carried out an undisclosed attack on RubyGems
Mecka AI nears $500M valuation in Sequoia-led deal amid rush for robot training data
So you want to use OpenRouter?
Houthis used Anthropic to develop guided weapons
Can LLMs Engineer Their Own Agent Harness? ByteDance Seed’s HarnessDev Says Only 34 of 64 Changes Generalize
GPT-6 built this earth exploration site in 5 prompts
AI researchers debate how close we are to recursive self-improvement
Anthropic Adds Plugin Evals to Claude Code: 6 Grader Types, a No-Plugin Baseline, and a CI Gate for Skills
Y Combinator’s Garry Tan wants US open-weight AI labs to ‘distill’ frontier models, too
OpenAI’s feud with mathematicians is only escalating
Lawyer fined $5K over AI-hallucinated witnesses in a murder case
One week left to book your exhibit table at TechCrunch Disrupt 2026
Final, final, final call for TechCrunch Disrupt 2026 Side Events
HotON.ai Radar filters noise from the AI market and highlights the changes that may affect cost, access, capability and competition.
View all →HotON.ai delivers market intelligence as video, visuals and text. Choose how you consume the AI economy — the platform handles every format.
A 90-second video recap of the week's biggest moves in AI prices, models and infrastructure.
Watch recapWhere input and output costs are rising and falling, at a glance.
Regional compute supply loosened this week, pushing the Inference Cost Index to a new monthly low across three major regions…
Showreels, recaps and explainers with adaptive playback.
Charts, infographics and covers, responsive and crisp.
Structured articles, summaries and data notes.
Structured intelligence on model pricing, AI infrastructure, inference cost, market heat and global AI supply-chain trends.
AI market briefings, pricing reports and index updates — in your inbox.
Access structured AI market data through HotON.ai APIs, feeds and custom intelligence products.
> GET /v1/models/OPUS-4.8/price { "symbol": "OPUS-4.8", "provider": "Anthropic", "input_per_1m": 6.00, "output_per_1m": 22.50, "context_k": 500, "efficiency": 96, "arena_elo": 1432, "modalities": ["text", "image"] }
HotON.ai is an AI market-intelligence platform tracking current prices (updated daily), token costs, quality (Arena Elo) and version price trends for 537 AI models across 79 providers worldwide.
Prices come from each provider's official pricing (via the TestKey catalog), cross-checked against OpenRouter, and refreshed daily. Every model page shows the price's source and 'as of' date.
Yes. Model prices, context windows, modalities and Arena Elo scores are real and sourced; efficiency and cost-per-task are computed from those real inputs and labeled as such.
As of Sep 12, 2026, the lowest blended price we track is $0.01 per 1M tokens. See the full 'cheapest models' ranking for the current list.
By LMArena human-preference Elo, the top-rated model we track is claude-opus-4.6 (1505). See the quality ranking for the full leaderboard.
It is a composite of a model's real price and context window, normalized to 0-100 so cheaper, larger-context models score higher. Full details are on our Methodology page.
HotON.ai gives you the data, indexes and intelligence to understand where the AI economy is moving next.