AI Chat for Work
Chat with multiple AI models to find the best output.
Upload files, do research, use domain expert skills.
45+ LLMs supported
Including the top open source models
Claude 3.5 SonnetClaude 3.5 SonnetClaude 3.7 SonnetClaude 3.7 SonnetClaude 3.7 Sonnet (T)Claude 3.7 Sonnet (Thinking)Claude 3.5 HaikuClaude 3.5 HaikuClaude 4 SonnetClaude 4 SonnetClaude 4 SonnetClaude 4 SonnetClaude 4 OpusClaude 4 OpusClaude 4 Opus (T)Claude 4 Opus (Thinking)Claude 4.5 OpusClaude 4.5 OpusClaude 4.6 OpusClaude 4.6 OpusClaude 4.5 SonnetClaude 4.5 SonnetClaude 4.5 SonnetClaude 4.5 SonnetClaude 4.5 HaikuClaude 4.5 HaikuClaude 4.5 HaikuClaude 4.5 Haiku
GPT-4oGPT-4oGPT-4o miniGPT-4o miniGPT-4.1GPT-4.1GPT-4.1 miniGPT-4.1 minio3o3o3 minio3 minio4 minio4 miniGPT-5GPT-5GPT-5 MiniGPT-5 MiniGPT-5 NanoGPT-5 NanoGPT-5.1GPT-5.1GPT-5.2GPT-5.2
Gemini 3 ProGemini 3 ProGemini 2.5 ProGemini 2.5 ProGemini 2.5 FlashGemini 2.5 FlashGemini 2.5 FlashGemini 2.5 FlashGemini 2.5 Flash LiteGemini 2.5 Flash LiteGemini 2.5 Flash LiteGemini 2.5 Flash LiteGemini 2.0 FlashGemini 2.0 Flash
Grok 4.1 FastGrok 4.1 FastGrok 4 FastGrok 4 FastGrok 4Grok 4Grok 3Grok 3Grok 3 MiniGrok 3 MiniGrok 2 VisionGrok 2 Vision
DeepSeek V3.2 SpecialeDeepSeek V3.2 SpecialeDeepSeek V3.2DeepSeek V3.2DeepSeek V3.1DeepSeek V3.1DeepSeek V3 (Chat)DeepSeek V3 (Chat)DeepSeek R1DeepSeek R1
Mistral Medium 3.1Mistral Medium 3.1Mistral Large 2.1Mistral Large 2.1Magistral Medium 1.2Magistral Medium 1.2Pixtral LargePixtral LargeMistral SabaMistral Saba
Perplexity SonarPerplexity SonarPerplexity Sonar ProPerplexity Sonar ProPerplexity Sonar Reasoning ProPerplexity Sonar Reasoning Pro
Kimi K2Kimi K2Kimi K2Kimi K2Kimi K2.5Kimi K2.5
Qwen 3 MaxQwen 3 MaxQwen 3 VL 235BQwen 3 VL 235BQwen 3 235B InstructQwen 3 235B InstructQwen 3 Coder PlusQwen 3 Coder PlusQwen 2.5-VL 32B InstructQwen 2.5-VL 32B Instruct
GLM 4.5GLM 4.5GLM 4.6GLM 4.6GLM 4.7GLM 4.7
Cogito v2 Llama 405BCogito v2 Llama 405BCogito v2 DeepSeek 671B MoECogito v2 DeepSeek 671B MoECogito v2.1 671BCogito v2.1 671BGPT OpenSource 20BGPT OpenSource 20BGPT OpenSource 120BGPT OpenSource 120B
MiniMax M2MiniMax M2MiniMax M2-herMiniMax M2-herFind your perfect output
Four ways to get better answers from AI.
Compare models. Pit Claude against GPT against Gemini in seconds. Pick the best part of each response.
Compare prompts. Tweak your wording until you find the right way to unlock dramatically better answers.
Research mode. Send up to 5 models in parallel to search the web and do deep-thinking for powerful research analyses.
Custom mode. Ready to race? Full manual control for every setting each model offers - temperature, thinking effort, and more.
Share, learn, improve
AI gets better when you don't go it alone.
Share your work. Found a killer prompt? Discovered which model aces your use case? One click generates a shareable link for teammates, clients, or your future self. Privacy-first—revoke access anytime.
Clone and remix. Every shared evaluation can be duplicated and tweaked. Learn from how others phrase their requests, then make it your own.
Prompt Doctor. Stuck? Our AI assistant analyzes what's not working and suggests improvements—like having an expert prompt engineer on call, 24/7.
Watch the demo
A full walkthrough of PromptHQ and how you can use it.
Frequently asked questions
Have more questions or just wanna talk prompts and evals? Email us at help@prompthq.run
What is Multichat?
Multichat lets you chat with up to 5 AI models simultaneously in side-by-side columns. Send one message, see how different models respond, and pick the best parts of each. It's like getting multiple expert opinions — you can cross-validate answers and combine the strongest ideas.
What models do you support?
We support 45+ models across all major providers — Anthropic (Claude), OpenAI (GPT, o-series), Google (Gemini), xAI (Grok), DeepSeek, Mistral, Perplexity, Kimi, Qwen, and more. New models are added regularly.
Why should I use PromptHQ over ChatGPT or Gemini?
ChatGPT and Gemini lock you into one provider. PromptHQ gives you all the top models in one place — premium and open source — so you can compare them side-by-side in Multichat, run structured Evals across models and prompts, and use Skills to get better results. You see which model actually works best instead of guessing.
Is it free?
Yes — the free plan includes lightweight models like GPT 5 Mini and Gemini 3 Flash, open source models like DeepSeek and Qwen, Skills, and most evaluation types. Paid plans ($10/mo and $25/mo) unlock premium models, web search, file uploads, and higher usage limits.
What are Skills?
Skills are specialized AI capabilities that help models perform specific tasks better — like writing SEO content, drafting documents, or analyzing data. Pick a skill, and every model in your chat benefits from that expertise.
Is my information private?
Yes. We only use the information you submit to serve and improve the product. We will not sell or distribute your data. You can share your data with people you choose through our collaboration features.
Do you train your LLM on my data?
No. We don't run our own LLM — we route your requests to the providers you choose (Anthropic, OpenAI, Google, etc.). Your data belongs entirely to you.
What are Evals?
Evals let you systematically compare models, prompts, and settings. Run the same prompt across multiple models, tweak temperature or thinking budgets, and see structured results — so you can make data-driven decisions instead of relying on vibes.
Who is the team behind PromptHQ.run?
PromptHQ.run is built by the team at GnoDesk. We built Multichat and Evals as internal tools to get better at using AI, and decided to make them publicly available.