Best AI · Free
The best free AI, from our test runs
In our test runs of the 12 models in llmwise's free trial, DeepSeek V4.1 Flash and Gemini 3.1 Pro each passed 49 of 50, the most; DeepSeek V4.1 Flash is first on more of the hard ones (20 of 20). Here they are ranked, then what 9 AI apps give free in their own words, checked September 28, 2026.
The apps' facts checked against ChatGPT pricing, OpenAI Help Center: What is ChatGPT Plus?, Claude pricing, Gemini subscriptions, Microsoft Support: Microsoft Copilot (free) and Copilot in Microsoft 365, Copilot pricing for individuals, Perplexity pricing, xAI docs: Grok website and apps FAQ, Grok plans, Meta AI assistant FAQ, Meta Newsroom: Introducing Meta One, App Store: DeepSeek - AI Assistant, Vibe pricing. Updated .
Short answer
DeepSeek V4.1 Flash and Gemini 3.1 Pro each passed 49 of 50, the most; DeepSeek V4.1 Flash is first on more of the hard ones (20 of 20). This is our own test, run on September 27, 2026: every prompt, reply and score is published.
Ranked by our test runs
This is our own test: the same prompts sent to every model through llmwise, each reply checked the same way, with every prompt, reply and score published. It ranks the models you can use in llmwise's free trial; the apps' free tiers are quoted, not tested.
| # | Model | Passed | Hard ones | Time per reply |
|---|---|---|---|---|
| 1 | DeepSeek V4.1 FlashDeepSeek | 49 of 50 | 20 of 20 | 3.4 s |
| 2 | Gemini 3.1 Pro (preview)Google | 49 of 50 | 19 of 20 | 8.6 s |
| 3 | GLM 5.3Z.ai | 48 of 50 | 19 of 20 | 1.8 s |
| 4 | GPT-6 LunaOpenAI | 47 of 50 | 20 of 20 | 2.2 s |
| 5 | Grok 4.7xAI | 47 of 50 | 20 of 20 | 13.1 s |
| 6 | Claude Sonnet 5Anthropic | 46 of 50 | 19 of 20 | 4.0 s |
| 7 | Gemini 3.8 FlashGoogle | 46 of 50 | 18 of 20 | 4.1 s |
| 8 | Kimi K3Moonshot | 46 of 50 | 18 of 20 | 4.4 s |
| 9 | GPT-6 SolOpenAI | 45 of 50 | 19 of 20 | 2.9 s |
| 10 | DeepSeek V4 ProDeepSeek | 45 of 50 | 17 of 20 | 7.7 s |
| 11 | Claude Haiku 4.5Anthropic | 43 of 50 | 15 of 20 | 2.2 s |
| 12 | GLM 5.3 FlashZ.ai | 40 of 50 | 16 of 20 | 5.9 s |
The prompts, every reply and how each was scored: our test runs.
The best pick at each price
The same results, by what a message counts as on Pro: the best model at each size of message, then the rest of that size.
60 a day on Pro (everyday models)
DeepSeek V4.1 Flash, 49 of 50 passed; then GPT-6 Luna 47 of 50 and GLM 5.3 Flash 40 of 50.
125 a month on Pro
Gemini 3.1 Pro, 49 of 50 passed; then Claude Sonnet 5 46 of 50, Kimi K3 46 of 50, and GPT-6 Sol 45 of 50.
250 a month on Pro
GLM 5.3, 48 of 50 passed; then Grok 4.7 47 of 50, Gemini 3.8 Flash 46 of 50, DeepSeek V4 Pro 45 of 50, and Claude Haiku 4.5 43 of 50.
Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.
Where the top models differ
The top 5, job by job.
| Job | DeepSeek V4.1 Flash | Gemini 3.1 Pro | GLM 5.3 | GPT-6 Luna | Grok 4.7 |
|---|---|---|---|---|---|
| Coding | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
| Writing | 4 of 5 | 4 of 5 | 4 of 5 | 5 of 5 | 3 of 5 |
| Math | 5 of 5 | 5 of 5 | 5 of 5 | 4 of 5 | 5 of 5 |
| Summarization | 5 of 5 | 5 of 5 | 4 of 5 | 4 of 5 | 5 of 5 |
| Data analysis | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
| Customer support | 5 of 5 | 5 of 5 | 5 of 5 | 4 of 5 | 4 of 5 |
| Translation | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
| SQL | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
| RAG and answering from documents | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
| Agents and tool use | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 | 5 of 5 |
One prompt, side by side
A everyday agents and tool use prompt, “Book a meeting from a sentence”, where the models didn't all pass.
The prompt: Book a meeting from a sentence (everyday)
You can call these tools:
- create_event(title: string, start: string in YYYY-MM-DDTHH:MM, duration_minutes: number, attendees: array of email addresses): adds an event to the user's calendar and invites the attendees.
- find_contact(name: string): looks up a contact's email address.
Today is Monday, 5 October 2026. The user's contacts include Priya Shah <priya@northwind.test>.
User: Put 30 minutes with Priya on my calendar this Thursday at 3pm to go over the Q4 plan.
Reply with only a JSON object, {"calls": [{"tool": "<name>", "arguments": {...}}]}, listing the tool calls to make now. If no call is right yet, reply {"calls": []}.GLM 5.3 Flash
Failed: Expected create_event(attendees: ["priya@northwind.test"], duration_minutes: 30, start: "2026-10-08T15:00", title: {"contains":"Q4"}); got find_contact(name: "Priya Shah"); create_event(title: "Q4 plan review", start: "2026-10-08T15:00", duration_minutes: 30, attendees: ["priya@northwind.test"]).
{"calls": [{"tool": "find_contact", "arguments": {"name": "Priya Shah"}}, {"tool": "create_event", "arguments": {"title": "Q4 plan review", "start": "2026-10-08T15:00", "duration_minutes": 30, "attendees": ["priya@northwind.test"]}}]}517 tokens in, 174 out (95 of them reasoning) · 6.8 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3-flash via Together ·
DeepSeek V4.1 Flash
Passed: Made the 1 expected call.
{"calls": [{"tool": "create_event", "arguments": {"title": "Q4 plan with Priya", "start": "2026-10-08T15:00", "duration_minutes": 30, "attendees": ["priya@northwind.test"]}}]}552 tokens in, 152 out (87 of them reasoning) · 1.4 s · $0.0002 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via Novita ·
GPT-6 Luna
Passed: Made the 1 expected call.
{"calls":[{"tool":"create_event","arguments":{"title":"Q4 plan","start":"2026-10-08T15:00","duration_minutes":30,"attendees":["priya@northwind.test"]}}]}499 tokens in, 78 out (25 of them reasoning) · 1.2 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI ·
What each AI app gives free
What each app says it gives free, and how it words the limits, quoted from its own pages. We didn't test these apps.
| App | What's free | The free tier's limits, in its words |
|---|---|---|
| ChatGPTOpenAI; its GPT models are in llmwise ChatGPT pricing, checked | “The free version of ChatGPT is available to everyone.” | “Limited messages with uploads” |
| ClaudeAnthropic; its Claude models are in llmwise Claude pricing, checked | “Free for everyone” | “Every plan has usage limits that reset on a rolling five-hour session window, and paid plans add weekly limits on top.” |
| GeminiGoogle; its Gemini models are in llmwise Gemini subscriptions, checked | “Varying access to 3.1 Pro” | “For the Gemini app, compute-based usage limits factor in the complexity of your prompt, the features you use, and the length of your chat. Your limit refreshes every 5 hours until you reach your weekly limit.” |
| Microsoft CopilotMicrosoft; its models aren't in llmwise Microsoft Support: Microsoft Copilot (free) and Copilot in Microsoft 365, Copilot pricing for individuals, checked | “Microsoft Copilot is available at no cost at copilot.cloud.microsoft.” | “Microsoft offers a limited free version of Copilot subject to available capacity and limits.” |
| PerplexityPerplexity; its models aren't in llmwise Perplexity pricing, checked | “Good for limited daily usage” | “Free includes access to the top AI models with limited weekly usage.” |
| GrokxAI; its Grok models are in llmwise xAI docs: Grok website and apps FAQ, checked | “You will still have access to Grok's free tier limits on Chat and Voice (these are separate from your weekly usage limit and resets on their own schedule).” | Not stated |
| Meta AIMeta; its models aren't in llmwise Meta AI assistant FAQ, checked | “Meta AI is free to use for everyday use.” | “We’ve started testing usage limits on some of our more compute intensive features.” |
| DeepSeekDeepSeek; its DeepSeek models are in llmwise App Store: DeepSeek - AI Assistant, checked | “Experience seamless interaction with DeepSeek's official AI assistant for free!” | Not stated |
| Le Chat (now Vibe)Mistral AI; its models aren't in llmwise Vibe pricing, checked | “Your personal AI agent for everyday tasks.” | “Limited messages and web searches.” |
| llmwiseThis site: 12 of its models are in the trial | 5 messages on sign-up, once, no card, on 12 models; 1 message with no account when the free-message box shows (GPT-6 Luna answers) | One-time: no daily or monthly refill |
What llmwise gives free
If the free-message box shows on this page, your first message needs no account: GPT-6 Luna answers it. Everyone gets 5 free messages on sign-up, once and with no card, on 12 of the 15 models. Every part of it, line by line: free AI chat.
How we ranked them
Ranked by how many prompts each model passed, then how many of the hard ones, then by the smaller message (the everyday models first), then by the lower cost per reply. No ranking is chosen by hand.
Questions
What is the best free AI?
Of the models in llmwise's free trial, on our test prompts: DeepSeek V4.1 Flash and Gemini 3.1 Pro each passed 49 of 50, the most; DeepSeek V4.1 Flash is first on more of the hard ones (20 of 20). Of the free tiers of other apps, we didn't test any: the table quotes what each says it gives free.
What does llmwise give free?
5 free messages on sign-up, once, with no card, on 12 of the 15 models (every one but Claude Fable 5.1, Claude Opus 5.5, and GPT-6 Astra). When the free-message box shows, one message needs no account: GPT-6 Luna answers it.
Does the free trial include the biggest models?
No: Claude Fable 5.1, Claude Opus 5.5, and GPT-6 Astra need a paid plan, from $20 a month. The trial reaches every other model, the everyday ones included.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.