Compare prompts across every model.
Side by side.

One prompt, every frontier model, one shareable link. GPT-5.4, Claude Opus 4.7, Gemini 3 Pro on day one, through a single key. No seats. No YAML.

Side-by-side comparisonShareable link

Prompt

How do I reset my password?

GPT-5.4

openai

Here's a 3-step plan to reset your password.

94

score

Claude Opus 4.7

anthropic

You can reset it from Settings → Account → Password.

97

score

Gemini 3 Pro

google

I'm not sure about that. Please contact support.

41

score

promptlens.io/r/q7k3m1p
Public

Every frontier model, one key. No per-provider signup.

OpenAI
Anthropic
Google
Meta
DeepSeek
Mistral
The Problem

Picking a model is guesswork.
Stakeholders review prompts in screenshots.

A prompt that works on Claude Opus 4.7 can tank on GPT-5.4. You find out by accident, after shipping. The tools you have turn every comparison into three browser tabs and a screenshot thread.

Without PromptLens

  • Open three tabs, paste the same prompt, squint at outputs
  • Sign up for each provider separately
  • Screenshot results into Slack to show a colleague
  • Lose a day figuring out which model actually answers best
  • Rerun everything the moment you tweak the prompt

With PromptLens

  • One view, every model, outputs aligned row by row
  • One API key via OpenRouter covers every provider
  • Share the comparison as a public link, no login
  • Deterministic scorers surface the winner automatically
  • Baseline + latest comparison catches regressions on the next run
Features

One prompt, every model, one link.

PromptLens is the comparison tool for developers picking models and sharing the answer.

Side-by-side comparison

Run the same prompt against every model you care about, outputs aligned row by row, scores on the right. No browser tab juggling.

Every model via one key

OpenAI, Anthropic, Google, Meta, DeepSeek, Mistral. GPT-5.4, Claude Opus 4.7, Gemini 3 Pro on day one. One key. No per-provider signup.

Shareable public reports

Every comparison gets a shareable link. Send it to your team, a stakeholder, or an investor. No account required to view, no per-seat pricing.

Deterministic scorers

Built-in metrics for similarity, keyword match, length range, regex, plus custom LLM-as-judge. Surfaces the winner without you eyeballing.

Baseline + latest diffs

Pin a baseline run, then every new run lines up against it. You see which outputs moved, which scores dropped, and which model drifted first.

Structured outputs, first class

Attach a JSON schema to any prompt. We validate every model's response and fail the cell if it doesn't conform. Great for agent tool-call comparison.

How it works

Three steps to a shareable comparison

90 seconds from blank page to a link your team can open.

1

Paste your prompt

Or pick a starter. Add a JSON schema if you want structured output validation.

2

Pick models to compare

One-click presets like GPT-5.4 vs Claude Opus 4.7 vs Gemini 3 Pro, or pick N models.

3

Share the link

Every run produces a public URL. Paste in Slack, a PR, or a deck. No login needed to view.

Built for picking models, not process compliance.

Heavier platforms sell you workspaces, RBAC, and tracing you don't need yet. PromptLens owns one job: give you the best answer to "which model should I use for this prompt" in one link.

No seats to buy
No provider keys to juggle
No screenshots in Slack
Side-by-side comparison
Run one prompt against N models in a single view. Outputs align row by row, scores on the right, diffs highlighted.
OpenRouter-unified access
Every frontier model through a single key. New releases like GPT-5.4 and Claude Opus 4.7 land the day OpenRouter picks them up, no config change on your side.
Shareable public reports
Slug-based public links. Optional password, optional expiry. Viewers don't need an account, you don't pay per seat.

One month from now

You'll stop opening three tabs to compare a prompt.

You'll pick the model by looking at the numbers, not by vibes.

You'll send one link instead of six screenshots.

Start your first comparison
Pricing

Simple, transparent pricing

Start free. Upgrade when you outgrow it. No seats, no surprise overage.

Free
$0
Best for trying out PromptLens
  • 3 projects
  • 50 evaluations/day
  • 3 share links
  • Every model, one key
Recommended
Pro
$99/month
For developers shipping LLM features
  • Unlimited projects
  • Unlimited comparisons
  • Unlimited share links
  • Baseline + latest diffs
  • Structured output validation
FAQ

Frequently asked questions

Common questions about PromptLens and how it compares to the tools you're already using.

Compare your prompt across every model.

No credit card. Shareable link in 90 seconds.