We test the hype. You get the truth — and the free alternatives.
AI tools change pricing and features faster than anyone can track. AI Tool Scout checks official sites, cuts through the marketing, and tells you what a tool really costs — and what you can use instead for free.
Pricing and limits are verified against official sources, not guesswork.
Reviews are refreshed as free trials vanish and prices change.
Every review answers the real question: what's the best free option?
Browse by Category
AI tools are grouped by what they do — and non-AI software has its own section so you always know what you're looking at.
💬 AI Chat Assistants 19 reviews
Meituan LongCat-2.5 Review (2026): 1.6T Params, 1M Context, No Weights
1.6 trillion parameters, about 48B active per token, a 1M context and $0.30/$1.20 per million tokens — a sixteenth of Opus 5.5 per token. Every benchmark is Meituan's own, and the weights are still a promise with no date.
Read the review →Claude Opus 5.5 Review (2026): Is AA 58 Worth $10 a Task?
58 on the independent index and 20% cheaper than Opus 5 at $4 in and $20 out per million tokens. At maximum effort the same third party measures $10.04 per task against $2.45 at medium, and Claude's text now carries a watermark you cannot see.
Read the review →Mistral Large 4 Review (2026): Open Weights, AA 26, $0.70/$2.10
1.05 trillion parameters with 52 billion active, a 1M context and $0.70 per million input tokens, with weights you can download. It scores 26 on the independent index: the best open model outside China, and still six points behind DeepSeek and GLM.
Read the review →Gemini 4 Carbon Review (2026): AA 59, $2/$10, No Access Yet
It ranks first on the independent intelligence index at 59 and costs $2 per million input tokens. On launch day only a subset of cyber defenders can reach it, Google published no context window and no model ID, and the price doubles to $4 / $20 with no end date.
Read the review →GPT-6.1 Terra Review (2026): $1/$5, Computer Use, 1.05M Context
$1 per million input and $5 per million output, a 1.05M-token context and computer use, at half the price of GPT-6.1 Sol. It sits 25th on the independent intelligence index, and prompts over 272K input tokens are billed at double rates.
Read the review →Kimi K3.1 Pro Review (2026): $2/$10, 2.8T, Weights Due Nov 4
A post-trained K3: 2.8T parameters, 104B active, a 1M-token context and $2/$10 per million tokens. It scores 55 on the independent intelligence index, but the weights are a 4 November promise and every benchmark is vendor-run.
Read the review →Inception Mercury 3 Review (2026): 1,068 Tokens/s, No Weights
A diffusion model that writes in parallel: 1,068 tokens per second, a 1M-token context and $0.10/$0.35 per million tokens at launch. Every benchmark is self-reported, and it still loses to Claude Haiku 5.5 on DeepSWE.
Read the review →MiniMax M3.1 Flash Review (2026): $0.15 Agents, No Weights
A 428B MoE with 23B active, a 1M-token context and flat $0.15/$0.60 per million tokens. MiniMax cut agent-task cost by 70% and says it is 3x faster, but every benchmark is vendor-reported and the weights are not out yet.
Read the review →DeepSeek V4.1-Flash-1017 Review (2026): Beats Opus 5 on Coding, Loses on Reasoning
A 763B model with 8B active parameters, a 1M-token context and MIT weights you can download. It scores 77.0 on DeepSWE against Claude Opus 5 at 74.0, at the same price as the September version. The reasoning gap is still wide.
Read the review →Meta Muse Spark Open Review (2026): The US Open-Weights Answer to Kimi K3
Meta calls it the most capable US open-weight model: 816B parameters, a 1M-token context and a gated download. It wins on tool use, but it loses to Kimi K3 on coding, and one independent agent benchmark puts it at just 8.1%. Here is what it really costs to run.
Read the review →Kimi K3 Review (2026): The Biggest Open-Weights Model & What It Really Costs
2.8 trillion parameters, a 1M-token context and the highest score of any open-weights model. The catch is 1.56TB of weights: open is not the same as runnable. Here is the API price, the real self-hosting bill, and who should use which.
Read the review →Gemini 4 Argon Review (2026): Pricing, Benchmarks & the November Price Hike
Google's new flagship tops the Vals Index and has the lowest hallucination rate measured — but independent scoring puts it behind Claude Opus 5.5, and the $2/$10 introductory price doubles on 30 November. Here is what it actually costs and where it actually wins.
Read the review →Grok Review (2026): Pricing, Free Plan & 5 Alternatives
Read the review →GPT-6 Astra Review (2026): Pricing, Benchmarks & Should You Switch?
Read the review →ChatGPT Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →Claude Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →Google Gemini Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →DeepSeek Review (2026): Free AI Chat, Reasoning & 5 Free Alternatives
Read the review →Qwen Review (2026): Free Chat, Cheapest AI API & 5 Free Alternatives
Read the review →🔍 AI Search & Research 1 review
Perplexity Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →🖼️ AI Image Generation 3 reviews
Ideogram Review (2026): Pricing, Free Plan & 5 Alternatives
Read the review →Midjourney Review (2026): Pricing, Free Trial & 5 Free Alternatives
Read the review →Leonardo AI Review (2026): Pricing, Free Tokens & 5 Free Alternatives
Read the review →🎬 AI Video & Avatars 5 reviews
Sora Shutdown (2026): What Happened & Best Free Alternatives
Read the review →Runway Review (2026): Pricing, Free Credits & 5 Free Alternatives
Read the review →Kling AI Review (2026): Pricing, Free Credits & 5 Free Alternatives
Read the review →Hailuo AI Review (2026): H3 Model, Pricing & 5 Free Alternatives
Read the review →HeyGen Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →🎙️ AI Voice & Music 3 reviews
Mistral Voxtral 3 Review (2026): Voice Agents at $0.04/min
Voice in, voice out, tool calls in between: three transcription tiers from $0.0035 a minute, TTS at $13 per million characters and a LiveKit agent SDK at $0.04 a minute. The realtime model still trails Gemini on word error rate.
Read the review →ElevenLabs Review (2026): Pricing, Free Tier & 5 Free Alternatives
Read the review →Suno Review (2026): Pricing, Free Songs & 5 Free Alternatives
Read the review →🪄 AI Photo & Face Swap 1 review
Unboring AI Review (2026): Pricing, Features & 5 Free Alternatives
Read the review →⌨️ AI Coding 2 reviews
GitHub Copilot Review (2026): Pricing, Free Plan & vs Cursor
Read the review →Cursor Review (2026): Pricing, Free Plan & 5 Free Alternatives
Read the review →🎨 AI Design 1 review
Canva Review (2026): Pricing, Free Plan & AI Magic Tools
Read the review →🧮 Developer Cost Guides 1 guide
LLM API Token Costs Compared (2026): Cheapest Models & 7 Ways to Cut Your Bill
Read the guide →✍️ AI Writing & Grammar 1 review
Grammarly Review (2026): Free Plan, Pricing & 5 Free Alternatives
Read the review →📝 AI Productivity 2 reviews
Otter.ai Review (2026): Pricing, Free Plan & 5 Alternatives
The free tier is a real product — 300 minutes a month with AI summaries and Zoom/Meet/Teams integration. But nothing is truly unlimited: even Pro caps file imports at 300 minutes per person, per month. Here is what you actually get, and 5 alternatives.
Read the review →Notion AI Review (2026): Pricing, Free Plan & 5 Alternatives
Read the review →🤖 AI Agents 3 reviews
Meta Muse Wonder Review (2026): Pricing, Benchmarks & What You Actually Get
Meta's first true frontier model prices at ~$4/$20 per million tokens, but independent scoring puts it below GPT-6 Astra and Claude Opus 5.5 — and the strongest checkpoint, Wonder Full, has no API price at all. Here is what you actually get, and where the benchmarks stop matching reality.
Read the review →OpenAI Dots Review (2026): Pricing, Free Plan & 5 Alternatives
OpenAI's always-on agents each get their own cloud computer and 4,000+ app connections — but the first dot only comes with a $100/month Pro plan. Free, Go and Plus are all excluded, and Pro dots are blocked in the UK, EEA and Switzerland. Here is what you actually pay, and 5 alternatives.
Read the review →Meta Muse Review (2026): Pricing, Free Plan & 5 Alternatives
Meta's personal agent hit #1 on both app stores and passed 3.4 million downloads in three weeks — but it is only available in the US and Canada, and the free tier reportedly still needs a card on file. Plus Muse Code at $5/month against $200 rivals, and 5 alternatives.
Read the review →💼 Software & SaaS Reviews 1 review
Non-AI business software — reviewed with the same data-checked standard.
Latest Reviews
Meituan LongCat-2.5 Review (2026): 1.6T Params, 1M Context, No Weights
1.6 trillion parameters, about 48B active per token, a 1M context and $0.30/$1.20 per million tokens — a sixteenth of Opus 5.5 per token. Every benchmark is Meituan's own, and the weights are still a promise with no date.
Read the review →Claude Opus 5.5 Review (2026): Is AA 58 Worth $10 a Task?
58 on the independent index and 20% cheaper than Opus 5 at $4 in and $20 out per million tokens. At maximum effort the same third party measures $10.04 per task against $2.45 at medium, and Claude's text now carries a watermark you cannot see.
Read the review →Mistral Large 4 Review (2026): Open Weights, AA 26, $0.70/$2.10
1.05 trillion parameters with 52 billion active, a 1M context and $0.70 per million input tokens, with weights you can download. It scores 26 on the independent index: the best open model outside China, and still six points behind DeepSeek and GLM.
Read the review →Gemini 4 Carbon Review (2026): AA 59, $2/$10, No Access Yet
It ranks first on the independent intelligence index at 59 and costs $2 per million input tokens. On launch day only a subset of cyber defenders can reach it, Google published no context window and no model ID, and the price doubles to $4 / $20 with no end date.
Read the review →GPT-6.1 Terra Review (2026): $1/$5, Computer Use, 1.05M Context
$1 per million input and $5 per million output, a 1.05M-token context and computer use, at half the price of GPT-6.1 Sol. It sits 25th on the independent intelligence index, and prompts over 272K input tokens are billed at double rates.
Read the review →Kimi K3.1 Pro Review (2026): $2/$10, 2.8T, Weights Due Nov 4
A post-trained K3: 2.8T parameters, 104B active, a 1M-token context and $2/$10 per million tokens. It scores 55 on the independent intelligence index, but the weights are a 4 November promise and every benchmark is vendor-run.
Read the review →Mistral Voxtral 3 Review (2026): Voice Agents at $0.04/min
Voice in, voice out, tool calls in between: three transcription tiers from $0.0035 a minute, TTS at $13 per million characters and a LiveKit agent SDK at $0.04 a minute. The realtime model still trails Gemini on word error rate.
Read the review →Inception Mercury 3 Review (2026): 1,068 Tokens/s, No Weights
A diffusion model that writes in parallel: 1,068 tokens per second, a 1M-token context and $0.10/$0.35 per million tokens at launch. Every benchmark is self-reported, and it still loses to Claude Haiku 5.5 on DeepSWE.
Read the review →MiniMax M3.1 Flash Review (2026): $0.15 Agents, No Weights
A 428B MoE with 23B active, a 1M-token context and flat $0.15/$0.60 per million tokens. MiniMax cut agent-task cost by 70% and says it is 3x faster, but every benchmark is vendor-reported and the weights are not out yet.
Read the review →DeepSeek V4.1-Flash-1017 Review (2026): Beats Opus 5 on Coding, Loses on Reasoning
A 763B model with 8B active parameters, a 1M-token context and MIT weights you can download. It scores 77.0 on DeepSWE against Claude Opus 5 at 74.0, at the same price as the September version. The reasoning gap is still wide.
Read the review →Meta Muse Spark Open Review (2026): The US Open-Weights Answer to Kimi K3
Meta calls it the most capable US open-weight model: 816B parameters, a 1M-token context and a gated download. It wins on tool use, but it loses to Kimi K3 on coding, and one independent agent benchmark puts it at just 8.1%. Here is what it really costs to run.
Read the review →Kimi K3 Review (2026): The Biggest Open-Weights Model & What It Really Costs
2.8 trillion parameters, a 1M-token context and the highest score of any open-weights model. The catch is 1.56TB of weights: open is not the same as runnable. Here is the API price, the real self-hosting bill, and who should use which.
Read the review →Gemini 4 Argon Review (2026): Pricing, Benchmarks & the November Price Hike
Google's new flagship tops the Vals Index and has the lowest hallucination rate measured — but independent scoring puts it behind Claude Opus 5.5, and the $2/$10 introductory price doubles on 30 November. Here is what it actually costs and where it actually wins.
Read the review →Meta Muse Wonder Review (2026): Pricing, Benchmarks & What You Actually Get
Meta's first true frontier model prices at ~$4/$20 per million tokens, but independent scoring puts it below GPT-6 Astra and Claude Opus 5.5 — and the strongest checkpoint, Wonder Full, has no API price at all. Here is what you actually get, and where the benchmarks stop matching reality.
Read the review →OpenAI Dots Review (2026): Pricing, Free Plan & 5 Alternatives
OpenAI's always-on agents each get their own cloud computer and 4,000+ app connections — but the first dot only comes with a $100/month Pro plan. Free, Go and Plus are all excluded, and Pro dots are blocked in the UK, EEA and Switzerland. Here is what you actually pay, and 5 alternatives.
Read the review →Meta Muse Review (2026): Pricing, Free Plan & 5 Alternatives
Meta's personal agent hit #1 on both app stores and passed 3.4 million downloads in three weeks — but it is only available in the US and Canada, and the free tier reportedly still needs a card on file. Plus Muse Code at $5/month against $200 rivals, and 5 alternatives.
Read the review →Otter.ai Review (2026): Pricing, Free Plan & 5 Alternatives
The free tier is a real product — 300 minutes a month with AI summaries and Zoom/Meet/Teams integration. But nothing is truly unlimited: even Pro caps file imports at 300 minutes per person, per month. Here is what you actually get, and 5 alternatives.
Read the review →Ideogram Review (2026): Pricing, Free Plan & 5 Alternatives
The best at rendering readable text inside images — but the free tier is now only ~10 slow credits per week, and everything you make is public. Here is what you actually get, and 5 alternatives.
Read the review →Notion AI Review (2026): Pricing, Free Plan & 5 Alternatives
Notion AI is now bundled into the Business plan (~$20/member/mo) — the standalone ~$8-10 add-on is gone. Here is what the free tier actually gives you, the credit system, and 5 alternatives.
Read the review →GPT-6 Astra Review (2026): Pricing, Benchmarks & Should You Switch?
OpenAI's new flagship costs 2.5x GPT-5.6 and needs a $100/mo plan for chat access — a real leap for agentic work, overpriced for everyday use.
Read the review →Grok Review (2026): Pricing, Free Plan & 5 Alternatives
xAI's Grok 4.6: a genuinely free tier, real-time X and web data, and API pricing from ~$2/$6 per 1M tokens — plus where it falls short of ChatGPT.
Read the review →GitHub Copilot Review (2026): Pricing, Free Plan & vs Cursor
Free tier with 2,000 completions a month, Pro from ~$10 — and the June 2026 switch to usage-based credits that heavy agent users should know about.
Read the review →Sora Shutdown (2026): What Happened & Best Free Alternatives
OpenAI shut down Sora in March 2026. Here's what happened and the genuinely free video generators to use instead.
Read the review →Qwen Review (2026): Free Chat, Cheapest AI API & 5 Free Alternatives
Alibaba's Qwen is one of the few truly free, no-ad AI chatbots — and its API is among the cheapest on the market.
Read the review →Grammarly Review (2026): Free Plan, Pricing & 5 Free Alternatives
The most famous writing assistant — the free plan is genuinely useful, Pro $12/mo. Here's what it really costs and the free alternatives.
Read the review →Runway Review (2026): Pricing, Free Credits & 5 Free Alternatives
Gen-4.5 is top-tier AI video — but the free tier is 125 one-time credits and Standard covers only ~52 seconds a month. Here's the real cost and the free alternatives.
Read the review →Wave Accounting Review (2026): Free Plan, Pricing & 5 Free Alternatives
The most genuinely free bookkeeping platform for small businesses — unlimited invoicing at $0. Here's what it costs and the free alternatives.
Read the review →Canva Review (2026): Pricing, Free Plan & AI Magic Tools
The easiest design tool on earth — free plan with daily AI credits, Magic tools built in. Here's what it costs and the free alternatives.
Read the review →Google Gemini Review (2026): Pricing, Free Plan & 5 Free Alternatives
The strongest fully free AI assistant — unlimited use, deep Google integration. Here's what it costs and the free alternatives.
Read the review →DeepSeek Review (2026): Free AI Chat, Reasoning & 5 Free Alternatives
The completely free assistant with top-tier reasoning and coding — no ads, no paywall. Here's why it matters.
Read the review →Claude Review (2026): Pricing, Free Plan & 5 Free Alternatives
The best writer among AI assistants — generous free plan, 200K context, and Claude Cowork. Here's what it costs and the free alternatives.
Read the review →LLM API Token Costs Compared (2026): Cheapest Models & 7 Ways to Cut Your Bill
Who's cheapest per million tokens, and the 7 tricks that cut most AI bills 60-90%.
Read the guide →Free Tools Directory
Free AI Tools Directory (2026): 60+ Best Free AI Tools by Category
Hand-picked free AI tools for chatbots, images, photo editing, video, audio, productivity, coding and research — with honest notes on what each free tier really gives you.
Browse the directory →Handy free tools, no sign-up
Honest accounting software reviews: Wave, QuickBooks, Xero & free options.
🎮 Gaming Gear ReviewsEsports hardware decoded: gaming mice, keyboards, headsets & budget picks.
📄 PDF Merge OnlineCombine multiple PDF files into one — free, no upload, files stay on your device.
📷 ID Photo StudioPassport & visa photos with AI background changer, compressor & converter.
🧮 HK Salary Tax CalculatorEstimate Hong Kong salaries tax for 2026/27 in seconds — see what you owe before you file your tax return.
🎁 HK Benefits Quick CheckGovernment & benefits eligibility checker — OAA, CSSA, housing & more
📅 HK Calendar 香港月曆Public holidays, lunar dates & solar terms for 2020-2036 — bilingual, free, no sign-up.
★ Our own product 🏠 Home Expense BookkeepingFamily bookkeeping: income, spending, net worth and monthly bills. Data stays on your device. HK$98 one-time.
