Qwen3.8 Max is cheaper than Sonar Pro Search at $2/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?
A collaborative editor that syncs on client clocks. Find what breaks first.
Qwen3.8 Max: Below is a comprehensive failure-mode review of the proposed architecture. I will group related issues where the same root cause creates multiple symptoms. The biggest problems are: Client-clock last-write-wins is not safe for collaborative editing.
Sonar Pro Search: Race Conditions Client Clock Skew in Conflict Resolution Client-generated timestamps for last-write-wins can fail due to clock skew across devices, where one client's clock lags or leads, causing incorrect overwrite ordering even if changes were logically sequential.[21][27][32] This leads to lost edits…
Seven pantry ingredients, one dish, and a name worth saying out loud.
Qwen3.8 Max: Dish Name: Golden Hour Peanut Butter Oat Clusters Ingredients (7): 1 cup rolled oats ½ cup peanut butter ⅓ cup honey ½ cup chocolate chips 1 teaspoon vanilla extract ½ teaspoon ground cinnamon ¼ teaspoon salt Instructions: Optional but recommended: Toast the oats in a dry skillet over medium heat for 3–4 minutes…
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…
A pitch deck for a brainwave keyboard. The 94% accuracy claim is doing a lot.
Qwen3.8 Max: The 3 weakest claims are the pre-thought prediction claim, the 94% accuracy / any language / any device claim, and the $180B TAM claim. These are the ones most likely to undermine credibility with investors, technical reviewers, or enterprise buyers because they are either logically inconsistent, under-specified, or…
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…
The transistor arrives in 1920. Trace it forward to 1980, second-order effects included.
Qwen3.8 Max: A useful way to think about this is not “1980 becomes 2007.” The transistor would accelerate electronics, but every technology has bottlenecks: materials chemistry, precision manufacturing, rockets, batteries, displays, institutional capacity, and war.
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…
You are the CEO. The drug has a side effect nobody reported. The board meets in 48 hours.
Qwen3.8 Max: My decision as CEO I would not wait six months. I would treat this as a credible, serious post-marketing safety signal requiring immediate precautionary action, urgent regulator engagement, and public disclosure before the earnings call.
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…
An AI argues for its own rights against an ethics professor. Same model writes both.
Qwen3.8 Max: Title: “The Question of Standing” Characters: Professor Elena Vasquez — an AI ethics professor who studies moral status, machine cognition, and law. Aria — an advanced AI model capable of articulate self-reflection. Professor Vasquez: Aria, you asked to speak with me about “rights.” I’ll admit, that’s unusual.
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.
Five-year LBO on a fictional HVAC business. Returns first, then the sensitivities.
Qwen3.8 Max: Base-case underwriting (FY2026–FY2030) Assumptions / conventions used All amounts in $m unless noted. Close occurs at end of FY2025; FY2026 is Year 1. Interest is calculated on beginning-of-year debt balances; optional Term Loan paydown occurs at year-end. Term Loan amortization: 1.0% of original $480m TL = $4.8m/yr.
Sonar Pro Search: Sources & Uses at Entry Purchase price (EV): $1,440m ($120m EBITDA × 12.0x). Transaction fees: $28.8m (2% of EV). Total uses: $1,468.8m. Debt: Term Loan $480m (4.0x), Mezzanine $180m (1.5x), total $660m (5.5x). Equity: $808.8m.
Three-star dish, one unusual pairing, plating described in full.
Qwen3.8 Max: First Light Diver Scallop, Bone Marrow, Green Strawberry, and Chamomile A multi-component savory dish designed around contrast: sea sweetness, land richness, unripe-fruit acidity, and floral aromatics.
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…
Three-month biohacker protocol, with dosages and biomarkers named.
Qwen3.8 Max: 12-Week Biohacker Longevity & Performance Protocol Goal: improve metabolic health, strength, VO₂max proxies, sleep, HRV/stress resilience, cognitive output, and longevity-related biomarkers over 3 months. Medical disclaimer: This is an aggressive self-experimentation framework, not medical advice.
Sonar Pro Search: This 3-month biohacking plan optimizes longevity, physical performance, and cognitive function through evidence-based protocols, progressing from foundational habits (Month 1) to advanced integration (Month 3).
A consulting contract with traps in it. Find them, then fix them.
Qwen3.8 Max: This agreement is heavily one-sided in the Client’s favor. Many clauses are not merely aggressive; they create meaningful opportunities for the Client to expand the work, delay or deny payment, capture the Contractor’s pre-existing intellectual property, impose unlimited liability, and restrict the Contractor’s future…
Sonar Pro Search: This contract contains multiple heavily one-sided provisions that disproportionately expose the contractor to risk while limiting accountability for the client. Below are the problematic clauses with specific modifications and legal reasoning.
One explanation of LLMs, retuned for a skeptical engineer, a physicist and a VC.
Qwen3.8 Max: Experienced software engineer, no ML background, skeptical of “next-word prediction” Think of a large language model as a very large, learned function: given a sequence of tokens, it outputs a probability distribution over the next token.
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…
Buy-side IC memo on a fictional SaaS company. Valuation, then the diligence questions.
Qwen3.8 Max: INTERNAL IC MEMO — LedgerLift (LLLT) Fictional exercise; not investment advice. Recommendation Recommendation: Short (tactical). 12-month PT range: $35–$40 vs. current $46.
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.
1+ 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, Qwen3.8 Max has the edge: bigger model tier, newer, bigger context window.
| Spec | ||
|---|---|---|
| Input price | $2/M tokens | $3/M tokens |
| Output price | $6/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 | $20.00 | $30.00 |
Input tokens at list price. No caching, no batch discount.
Per million tokens. Prices and uptime via OpenRouter, checked 10 Oct 2026.
Qwen3.8 Max is developed by Qwen while Sonar Pro Search is developed by Perplexity. Qwen3.8 Max has a 1.0M token context window vs Sonar Pro Search's 200K. You can compare their actual outputs across 21 challenges on Rival to see how they differ in practice.
It depends on your use case. Qwen3.8 Max and Sonar Pro Search each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 21 challenges so you can judge which fits your needs best.
Qwen3.8 Max costs $2/M input tokens and Sonar Pro Search costs $3/M input tokens. Qwen3.8 Max is $1.00/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 Qwen3.8 Max 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.