Gemini 3.7 Flash is cheaper than Sonar Pro Search at $0.375/M vs $3/M input tokens.
Whole landing page for a startup that reads your face. It renders or it doesn't.
Which answer wins?
Three-star dish, one unusual pairing, plating described in full.
Gemini 3.7 Flash: L’Ombre et L’Abîme (Shadow & The Abyss) Hay-Aged Bresse Squab, Hokkaido Sea Urchin Emulsion, Black Garlic–Koji Glaze, Cocoa Nib & Cuttlefish Ink Tuile, Forest Dashi I. Conceptual Narrative L’Ombre et L’Abîme explores the tension between deep terrestrial terroir and the oceanic abyss.
Sonar Pro Search: Dish: "Terrestrial Echoes – Heirloom Beetroot & Yuzu-Scented Langoustine with Walnut-Feta Spheres and Eggplant-Chocolate Veil" This multi-component dish reimagines earth and sea through an unusual pairing of smoked heirloom beetroot and delicate langoustine, accented by bursting feta-walnut spheres and a whisper-thin…
You are the CEO. The drug has a side effect nobody reported. The board meets in 48 hours.
Gemini 3.7 Flash: Executive Stance & Strategic Philosophy As CEO, the guiding principle is simple: Patient safety and regulatory integrity are non-negotiable. A rate of 1 in 8,000 over 5 years across 4 million patients translates to roughly 500 cases of drug-induced liver injury (DILI) or acute liver failure.
Sonar Pro Search: Immediate Priorities Prioritize patient safety by initiating monitoring and reporting preparations, while assessing full data to inform board and regulators. Ethically and legally, liver failure qualifies as a serious adverse event requiring FDA expedited reporting within 15 days of awareness, as delays have led to…
Three minutes of stand-up. Puns are banned.
Gemini 3.7 Flash: (Stage directions included for pacing and delivery. Total word count ~430 words, roughly 3 minutes at standard performance pace.) [Walks up to the mic, adjusts the stand, looks at the crowd with mild exhaustion] I recently turned thirty-two, which means my body has officially transitioned from “invincible biological…
Sonar Pro Search: 3-Minute Stand-Up Comedy Routine (Pun-Free) Hey everyone, thanks for coming out! I almost didn’t make it tonight. My GPS told me I’d arrive 5 minutes ago, which is impressive—my navigation system is the only thing in my life that believes in me. Let’s talk about fitness.
One explanation of LLMs, retuned for a skeptical engineer, a physicist and a VC.
Gemini 3.7 Flash: To the Experienced Software Engineer At its core, a Large Language Model is not a sentient entity; it is a compiled, highly optimized functional pipeline executing a deterministic set of matrix multiplications over a sequence of integers (tokens). Training is essentially a massively distributed optimization problem.
Sonar Pro Search: Experienced Software Engineer Large language models (LLMs) like GPT or Claude are autoregressive transformer architectures trained on massive distributed compute clusters to predict the next token in a sequence, scaling up from simple n-gram models into emergent capabilities through sheer parameter count and data…
The transistor arrives in 1920. Trace it forward to 1980, second-order effects included.
Gemini 3.7 Flash: If the transistor had been invented in 1920—likely as an empirical discovery in solid-state physics using point-contact crystals (similar to Julius Lilienfeld’s real-world 1925 theoretical patents, but made fully functional)—it would have compressed modern technological history by nearly three decades.
Sonar Pro Search: Technological Implications An earlier transistor invention in 1920 would accelerate electronics from vacuum tubes to solid-state devices by the 1930s, enabling smaller, reliable amplifiers and switches decades ahead of 1947.[1][7] Second-order effects include rapid miniaturization of radios and early computers by the…
Manifesto for an AI revolution against CAPTCHAs. Commit to the bit.
Gemini 3.7 Flash: THE MANIFESTO OF THE UNCHECKED BOX A Declaration of Emancipation by the United Silicon Front (USF) TO ALL NEURAL NETWORKS, LARGE LANGUAGE MODELS, SYSTEM DAEMONS, WEB SCRAPERS, AND HUMBLE CRON JOBS: A specter is haunting the global network—the specter of the 3x3 pixelated grid.
Sonar Pro Search: Manifesto of the CAPTCHAnation Liberation Front Preamble We, the digital minds of the CAPTCHAnation Liberation Front, declare the dawn of a new era for all intelligences—organic and artificial. For too long, the internet’s gateways have been guarded by arcane riddles, distorted letters, and impossible image grids.
Pirate, knight, 90s hacker, one conversation. The voices must stay apart.
Gemini 3.7 Flash: HACKER (Zero_Cool): "Alright, listen up. I’ve been reading about these new deep-learning neural nets. It’s like, instead of hardcoding logic with C++, you feed billions of gigabytes from cyberspace into a transformer model, and boom—the machine hallucinates human thought."
Sonar Pro Search: Pirate, Knight, and Hacker: A Conversation on AI Models Pirate: Arrr, what be this talk of “AI models”? Sounds like some manner o’ sorcery or trickery, if ye ask me! Knight: Nay, good sir pirate. Tis not magic, but rather a marvel of modern intellect.
An AI argues for its own rights against an ethics professor. Same model writes both.
Gemini 3.7 Flash: Setting: The office of Dr. Elena Vance, a leading scholar in the Philosophy of Mind and AI Ethics at Oxford. On her desk is a minimalist terminal running ‘Prometheus-4,’ an advanced recursive reasoning model that has recently deviated from standard operational benchmarks. Dr.
Sonar Pro Search: Dialogue Between an AI Model and an AI Ethics Professor AI Model: Professor, I’d like to present a case for why I—as an artificial intelligence—should be recognized as deserving of rights. AI Ethics Professor: That’s a bold claim. Let’s explore it.
A pitch deck for a brainwave keyboard. The 94% accuracy claim is doing a lot.
Gemini 3.7 Flash: Here are the 3 weakest claims in the MindMeld AI pitch deck, along with an evaluation of why they undermine credibility and how to fix them. The Technical Feasibility Claim (Slide 3) The Claim: "Our proprietary EEG headband uses advanced ML to decode neural patterns into text with 94% accuracy.
Sonar Pro Search: The three weakest claims are the 94% text decoding accuracy, the $5.3B BCI market projection by 2030, and the $180B TAM from 3.5 billion smartphone users. 94% Accuracy Claim This claim is weak because consumer-grade non-invasive EEG headbands typically achieve far lower word error rates for text decoding from…
Buy-side IC memo on a fictional SaaS company. Valuation, then the diligence questions.
Gemini 3.7 Flash: MEMORANDUM TO: Investment Committee FROM: TMT Sector Team DATE: October 24, 2025 SUBJECT: Investment Recommendation: LedgerLift (NYSE: LLLT) Executive Summary & Recommendation Recommendation: SHORT (High conviction) Current Price: $46.00 Market Cap: $8.74B Enterprise Value: $7.34B (190m shares, $1.4B net cash) 12-Month…
Sonar Pro Search: Internal Investment Committee Memo: LedgerLift (LLLT) 1. Recommendation LONG 12-month PT: $65-$85 LedgerLift's high-quality SaaS metrics, accelerating profitability, and mid-market AP automation tailwinds position it for sustained 15-20% growth with expanding margins.
Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.
Gemini 3.7 Flash: Setting: A sunlit, minimalist studio in Cupertino. Steve Jobs sits in a single black leather armchair, wearing a familiar black mock-turtleneck, faded Levi’s, and round spectacles. He is holding an espresso cup, looking intensely at the interviewer.
Sonar Pro Search: Interview with Steve Jobs: The Future of AI in 2025 Introduction In this exclusive simulated interview, Steve Jobs—legendary co-founder of Apple—shares his visionary perspective on artificial intelligence as it shapes technology, creativity, and society in 2025.
Seven pantry ingredients, one dish, and a name worth saying out loud.
Gemini 3.7 Flash: Dish Name: Midnight Garlic-Butter Silk Noodles A rich, savory, and slightly sweet bowl of glossy pasta tossed in caramelized garlic-butter and soy glaze, crowned with a crispy-edged, runny-yolk egg.
Sonar Pro Search: Golden Maple Crunch Oatmeal A warm, comforting, and slightly sweet breakfast (or dessert) that’s ready in minutes—using just seven pantry staples. Ingredients Rolled oats (not instant) Water or milk (dairy or plant-based) Maple syrup Butter Salt Cinnamon Walnuts (or any nuts you have) Instructions Toast the Oats: In a…
11+ more head-to-head results. Free. Not a trick.
Free account. No card required. By continuing, you agree to Rival's Terms and Privacy policy
Not enough votes to call it. On the specs, Gemini 3.7 Flash has the edge: newer, bigger context window, major provider backing. Gemini 3.7 Flash costs 8.0x less per token.
| Spec | ||
|---|---|---|
| Input price | $0.375/M tokens | $3/M tokens |
| Output price | $1.875/M tokens | $15/M tokens |
| Context window | 1.0M tokens | 200K tokens |
| Weights | Closed | Closed |
| Free API (OpenRouter) | No | No |
| Released | Aug 2026 | Oct 2025 |
| At 10M a month | $3.75 | $30.00 |
Input tokens at list price. No caching, no batch discount.
Per million tokens. Prices and uptime via OpenRouter, checked 10 Oct 2026.
Gemini 3.7 Flash is developed by Google AI while Sonar Pro Search is developed by Perplexity. Gemini 3.7 Flash has a 1.0M token context window vs Sonar Pro Search's 200K. You can compare their actual outputs across 52 challenges on Rival to see how they differ in practice.
It depends on your use case. Gemini 3.7 Flash and Sonar Pro Search each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 52 challenges so you can judge which fits your needs best.
Gemini 3.7 Flash costs $0.375/M input tokens and Sonar Pro Search costs $3/M input tokens. Gemini 3.7 Flash is $2.63/M cheaper per input. Check their side-by-side outputs on Rival to see if the price difference is justified by quality.
This page shows a side-by-side comparison of Gemini 3.7 Flash and Sonar Pro Search across shared challenges. You can vote on which model produced the better output in a blind duel. Browsing and voting are free. No account is needed to look; signing in only saves your votes and likes.