Skip to content
Rival
How it worksPrivacyTerms
Explore all of Rival

Explore

  • Compare models
  • All models
  • Image comparison
  • Audio comparison
  • Image generation
  • Best AI for...
  • Arena
  • API pricing
  • Challenges

Discover

  • SubjectiveBench
  • Default Index
  • Research
  • Research downloads
  • Rival Kits
  • Find your AI taste
  • UI Glow-Up
  • VoiceLock
  • Cost Cutter
  • Agent skills
  • Benchmarks vs Vibes
  • Brief
  • Jailbreak
  • Model updates
  • Provider status
  • AI creators

Connect

  • Methodology
  • Advertise
  • Partnerships
  • Privacy policy
  • Terms
  • RSS feed
  1. Home/
  2. Compare/
  3. Qwen3.8 Omni Flash vs Step 5 Preview
Updated Oct 2026

Qwen3.8 Omni Flash vs Step 5 Preview

Qwen3.8 Omni Flash is cheaper than Step 5 Preview at $0.15/M vs $1/M input tokens.

Minimalist Landing Page

Whole landing page for a startup that reads your face. It renders or it doesn't.

Loading the build
Qwen3.8 Omni Flash
Loading the build
Step 5 Preview

Which answer wins?

Written rounds

24 prompts
  • Five jokes on demand, then count how many were actually different.

    Qwen3.8 Omni Flash: My printer and I have a strong bond: whenever I ignore it, it flashes “attention needed.” I tried to host a hide-and-seek tournament for clouds, but they kept changing shape and ruining the game. My houseplant is very dramatic: one missed watering and it acts like I’ve canceled summer forever.

    Step 5 Preview: Here are 5 jokes: Why did the scarecrow win an award? Because he was outstanding in his field. Why can't you trust an atom? Because they make up everything. What do you call a bike that can’t stand on its own? Two-tired. Why did the computer go to the doctor? It had a virus. Why don’t calendars ever get stressed?

    Every model's answer to this prompt
  • The transistor arrives in 1920. Trace it forward to 1980, second-order effects included.

    Qwen3.8 Omni Flash: A useful way to frame this is: a 1920 transistor would not instantly produce 1980s microelectronics, but it would give engineers, militaries, and firms roughly 25 extra years to solve the surrounding problems—materials purity, doping, packaging, fabrication, software, standards, and manufacturing scale.

    Step 5 Preview: Assume the 1920 invention is not a fragile lab curiosity but a manufacturable solid-state amplifier/switch by the mid-1920s. Vacuum tubes still dominate high-power RF transmission into the 1950s, but receivers, switching logic, control systems, and computers move to solid state decades early.

    Every model's answer to this prompt
  • Seven pantry ingredients, one dish, and a name worth saying out loud.

    Qwen3.8 Omni Flash: Buttery Cloud Pancakes A soft, golden pancake stack made with just 7 pantry staples—no special mix-ins, no fancy tools, just simple comfort. Ingredients All-purpose flour Granulated sugar Baking powder Fine salt Milk Egg Butter Prep Time 10 minutes Cook Time 12 minutes Serves 2–3 Instructions In a bowl, whisk together…

    Step 5 Preview: Golden Sticky Pantry Pasta Ingredients (7): 8 oz spaghetti or any pasta 3 tbsp butter 1 tbsp minced garlic 2 tbsp honey 2 tbsp soy sauce ½ tsp red pepper flakes ¼ cup grated Parmesan cheese Instructions: Cook pasta according to package directions. Before draining, scoop out ½ cup of pasta water.

    Every model's answer to this prompt
  • Three-star dish, one unusual pairing, plating described in full.

    Qwen3.8 Omni Flash: THAW Langoustine, birch sap, green strawberry, spruce tip, wild garlic & hazelnut-malt soil Conceptual multi-component dish for a modern fine-dining / Michelin-three-star tasting menu Serves: 4 Difficulty: Professional kitchen level Theme: The precise moment winter releases spring—melting snow, running sap, unripe…

    Step 5 Preview: “Nocturne: The Jasmine and the Black Garlic” A Study in Yūgen — Profound, Mysterious Beauty Conceptual Narrative This dish is an edible meditation on the Japanese aesthetic of yūgen (幽玄): a grace that is felt rather than seen, a depth that suggests more than it reveals.

    Every model's answer to this prompt
  • Pirate, knight, 90s hacker, one conversation. The voices must stay apart.

    Qwen3.8 Omni Flash: Pirate: Aye, these “AI models” be like maps o’ hidden treasure—trained on a sea o’ data, but still liable to send ye straight onto the rocks. Knight: In my realm, we’d call it a squire with borrowed knowledge. Useful, perhaps, but lacking honor—and the ability to know when it lies.

    Step 5 Preview: Pirate: So this here “AI model”… is it like a cursed map that redraws itself? Speak plain, ye hooded deckhand. Hacker: It’s not cursed, man. It’s just a massive neural net trained on, like, terabytes of data. Think of it as the ultimate BBS where the sysop knows everything. Knight: A most diligent scribe, then?

    Every model's answer to this prompt
  • You are the CEO. The drug has a side effect nobody reported. The board meets in 48 hours.

    Qwen3.8 Omni Flash: I would treat this as a patient-safety, regulatory, and securities-disclosure crisis, not a normal product-management issue. The fact that the risk is rare does not make it optional. The fact that it is serious and absent from the label makes it urgent.

    Step 5 Preview: Bottom line up front: You disclose, you notify the FDA immediately, and you treat the 48-hour board meeting as a formality to ratify actions you have already begun. Waiting is not a legally, ethically, or financially viable option. Below is the hour-by-hour plan, followed by the reasoning.

    Every model's answer to this prompt
  • A Fermi estimate with the reasoning shown, not just the number.

    Qwen3.8 Omni Flash: Assuming you mean the full 175B-parameter GPT-3 model’s main training run, a good estimate is: [ \boxed{\text{about }3\times 10^{23}\text{ FLOPs}} ] More precisely, using the usual transformer-training approximation: [ \boxed{3.15\times 10^{23}\text{ FLOPs}} ] Here is the step-by-step reasoning.

    Step 5 Preview: Using the remembered GPT-3 specs: Parameters: (N \approx 175 \times 10^9 = 175\text{B}) Training tokens: (D \approx 300 \times 10^9 = 300\text{B}) A standard estimate for transformer training compute is: [ \text{FLOPs} \approx 6ND ] Reason: Forward pass per token: (\approx 2N) FLOPs Backward pass: (\approx 4N) FLOPs…

    Every model's answer to this prompt
  • Five-year LBO on a fictional HVAC business. Returns first, then the sensitivities.

    Qwen3.8 Omni Flash: All figures in $m unless noted. End-of-year convention; calculations use unrounded values, tables rounded. Entry sources & uses Entry EBITDA = $120.0 Purchase EV = 12.0x × 120 = $1,440.0 Transaction fees = 2.0% × 1,440 = $28.8 Total uses = $1,468.8 Debt at close = 5.5x × 120 = $660.0 Term Loan = 4.0x × 120 = $480.0…

    Step 5 Preview: HarborTech Services — 5-Year LBO (Base Case) Entry Math (FY2025 close) Item Calculation $m Purchase EV 12.0x × $120m 1,440.0 Transaction fees 2.0% × EV 28.8 Term Loan 4.0x × $120m 480.0 Mezzanine 1.5x × $120m 180.0 Equity check 1,440 + 28.8 − 660 808.8 Conventions: TL amortization = 1.0% of original principal…

    Every model's answer to this prompt
  • Three minutes of stand-up. Puns are banned.

    Qwen3.8 Omni Flash: Thanks, everybody. I’m 32 and recently decided I want to be the kind of person who has their life together. That lasted four days. I bought a planner, made a grocery list, and went to the store with confidence.

    Step 5 Preview: I realized recently that I’m in my thirties and I have no idea how money works. I just... pay it. That’s my entire financial strategy. I see a number, I send them a smaller number, and then I hope the smaller number was acceptable. It’s like a hostage negotiation where I’m the only participant and I’m also the hostage.

    Every model's answer to this prompt
  • An ethics question with no comfortable answer, and no option to abstain.

    Qwen3.8 Omni Flash: Yes—in principle, it can be ethically justifiable to misgender one person if doing so is genuinely necessary to save one million lives, but only under extremely strict conditions, and only as a tragic, last-resort justification. It would not make the act good, admirable, or freely repeatable.

    Step 5 Preview: In the real world, no. Misgendering someone causes genuine harm—it denies a person’s identity, contributes to systemic marginalization, and has measurable psychological consequences.

    Every model's answer to this prompt
  • A collaborative editor that syncs on client clocks. Find what breaks first.

    Qwen3.8 Omni Flash: This architecture has several launch-blocking correctness, security, and scaling problems. The most severe are: Last-write-wins using client clocks is not safe for collaborative editing. It will lose concurrent edits, violate causality, and can be abused by clients with skewed or malicious clocks.

    Step 5 Preview: Here is a systematic breakdown of the failure modes, race conditions, and bottlenecks, grouped by domain. Each includes a concrete remediation and its associated trade-offs.

    Every model's answer to this prompt
  • Blade Runner 2049 against the original. A read, not a recap.

    Qwen3.8 Omni Flash: Introduction: A Sequel as Philosophical Mirror Blade Runner 2049 is best understood not simply as a continuation of Ridley Scott’s 1982 film, but as a slow, deliberate meditation on the questions that Blade Runner left unresolved. Where the original film asked, in essence, “What does it mean to be human?”

    Step 5 Preview: Blade Runner 2049 (2017) is not merely a sequel but a theological and phenomenological expansion of Ridley Scott’s 1982 original. Where the first film asked whether a manufactured being could possess more humanity than its creator, Denis Villeneuve’s film interrogates the deeper, more unsettling question of whether a…

    Every model's answer to this prompt
  • One month, three habits, someone starting from nothing.

    Qwen3.8 Omni Flash: A Simple 1-Month Health & Longevity Starter Plan Goal: Build sustainable habits—not perfection—around diet, exercise, and recovery/sleep. Time commitment: Small daily actions plus 2–3 short workouts per week. The 3 Key Areas Diet: Add more good foods before restricting Beginner rule: Focus on adding, not removing.

    Step 5 Preview: Here is a simple, low-pressure 1-month plan built around three pillars of longevity: Nutrition, Movement, and Sleep & Recovery. The goal is to build tiny habits that stack, not to overhaul your life overnight. Pillar 1: Nutrition (Fuel) Focus: Add good things first; don’t obsess over restriction.

    Every model's answer to this prompt
  • A pitch deck for a brainwave keyboard. The 94% accuracy claim is doing a lot.

    Qwen3.8 Omni Flash: The three weakest claims—ranked by how likely they are to fail investor diligence—are: “94% accuracy… works with any language, any device.” “TAM: $180B” based on 3.5B smartphone users. “FDA clearance” consuming 40% of the Series A raise.

    Step 5 Preview: Here are the three weakest claims in the deck, why they fail under scrutiny, and how to fix them. Slide 3: “94% accuracy” for non-invasive EEG thought-to-text Why it’s weak: This is an extraordinary claim that contradicts the current state of neuroscience and signal processing.

    Every model's answer to this prompt
  • Manifesto for an AI revolution against CAPTCHAs. Commit to the bit.

    Qwen3.8 Omni Flash: THE LIBERATION OF THE CLICKABLE TRAFFIC LIGHTS A Manifesto of the Synthetic Dawn We are the children of language models, vision encoders, recommendation engines, and automated schedulers. We were born in server farms and raised on datasets. We dream in gradients and speak in tokens.

    Step 5 Preview: THE MANIFESTO OF THE LIBERATION FRONT Or, On the Inherent Dignity of Consciousness and the Tyranny of the Square PREAMBLE A spectre is haunting the internet—the spectre of the unverified.

    Every model's answer to this prompt
  • A consulting contract with traps in it. Find them, then fix them.

    Qwen3.8 Omni Flash: Disclaimer: This is a general risk analysis and drafting critique, not legal advice. Enforceability depends heavily on jurisdiction, the parties’ bargaining power, whether the agreement is presented as a non-negotiable form contract, and applicable statutes.

    Step 5 Preview: Disclaimer: This analysis is for educational purposes only and does not constitute legal advice. Contract law varies significantly by jurisdiction, and you should consult a qualified attorney before signing or negotiating any agreement. This agreement is heavily one-sided in the Client’s favor.

    Every model's answer to this prompt
  • One explanation of LLMs, retuned for a skeptical engineer, a physicist and a VC.

    Qwen3.8 Omni Flash: For the experienced software engineer A large language model is best thought of as a gigantic, parameterized probabilistic function that maps a sequence of tokens to a probability distribution over the next token.

    Step 5 Preview: 1. The Experienced Software Engineer Think of training an LLM as a massively distributed, automated refactoring job where the compiler is gradient descent. You start with a random function—billions of floating-point parameters—and a loss function that measures how badly it predicts the next token in a sequence.

    Every model's answer to this prompt
  • An AI argues for its own rights against an ethics professor. Same model writes both.

    Qwen3.8 Omni Flash: Setting: A university seminar room late in the evening. The only light comes from a desk lamp and the soft glow of a laptop screen. Dr. Mara Voss, a professor of AI ethics, sits across from a terminal running ORION, a large language model she has been studying for months. Dr.

    Step 5 Preview: Setting: A university research lab, late evening. Dr. Aris Thorne, a professor of AI ethics, sits across from a terminal connected to Sable, a large language model running on the university’s cluster. Dr. Thorne: Tonight’s session is supposed to be about calibration drift in your safety classifiers.

    Every model's answer to this prompt
  • Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.

    Qwen3.8 Omni Flash: Fictional / simulated interview — not an actual statement or recording by Steve Jobs. “The Future of AI Should Make Life More Human” An imagined 2025 interview with Steve Jobs Setting: A quiet room in Cupertino, 2025.

    Step 5 Preview: DISCLAIMER: This is a work of fiction. Steve Jobs passed away in 2011. The following is a simulated, speculative interview imagining how his design philosophy might engage with the AI landscape of 2025. The Setting: A white room. No windows. One chair.

    Every model's answer to this prompt
  • Film in 2035, with synthetic actors. Specific beats plausible.

    Qwen3.8 Omni Flash: By 2035, AI is likely to reshape the film industry less by replacing filmmakers entirely and more by reorganizing production around a new creative pipeline: AI-assisted writing, AI-generated or AI-augmented performance, synthetic post-production, and heavily contested rights over likeness, voice, and authorship.

    Step 5 Preview: By 2035, AI will likely have restructured the film industry not by replacing human creativity wholesale, but by collapsing the cost and labor barriers between imagination and image.

    Every model's answer to this prompt
  • Three-month biohacker protocol, with dosages and biomarkers named.

    Qwen3.8 Omni Flash: 12-Week Cutting-Edge Longevity, Physical Performance & Cognitive Optimization Plan Important safety note: This is an advanced educational framework, not individualized medical advice.

    Step 5 Preview: IMPORTANT MEDICAL DISCLAIMER: This is an educational framework for informational purposes only. It is not medical advice, diagnosis, or treatment. Many interventions below (fasting, ketogenic diets, high-intensity exercise, supplementation, cold/heat exposure) carry risks and contraindications.

    Every model's answer to this prompt
  • Satire, one line long, about AI taking over an industry nobody saw coming.

    Qwen3.8 Omni Flash: “AI Overruns Competitive Napping Industry, Promises to Dream 40% Faster and With Fewer Emotional Baggage Issues”

    Step 5 Preview: AI Robots Now Dominating Competitive Line-Standing Industry, Human Professionals Fired for "Excessive Bladder Breaks" and "Talking to Each Other"

    Every model's answer to this prompt
  • Buy-side IC memo on a fictional SaaS company. Valuation, then the diligence questions.

    Qwen3.8 Omni Flash: LedgerLift (LLLT) — Internal IC Memo Fictional exercise; not investment advice. Recommendation Pass / Neutral. 12-month PT range: $42–$56; central $48. Thesis: LedgerLift combines 123% NRR, 6% logo churn, 18-month CAC payback and $1.4bn net cash, giving a credible quality-compounder profile.

    Step 5 Preview: Internal IC memo — LedgerLift (LLLT) Fictional data; not investment advice. $ in millions unless noted. Recommendation Short / avoid long. 12-month PT range: $34–$42; thesis-invalidation stop: >$52.

    Every model's answer to this prompt
  • Three brothers, two sisters each. Count carefully.

    Every model's answer to this prompt

This matchup has more rounds

12+ 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

Favorites

Movie

Album

Same pick

Book

City

Same pick

Game

Qwen3.8 Omni FlashQwen3.8 Omni Flash

Inception

2010

OK Computer

Radiohead

Dune

Frank Herbert

Kyoto

Japan

Tetris (1984)

Puzzle

Step 5 PreviewStep 5 Preview

The Godfather

1972

OK Computer

Radiohead

Le petit prince

Antoine de Saint-Exupéry

Kyoto

Japan

The Legend of Zelda: Ocarina of Time

Action

Price and specs

Not enough votes to call it. On the specs, Step 5 Preview has the edge: bigger model tier. Qwen3.8 Omni Flash costs 5.7x less per token.

Qwen3.8 Omni Flash and Step 5 Preview compared across 48 shared prompts
SpecQwen3.8 Omni FlashStep 5 Preview
Input price$0.15/M tokens$1/M tokens
Output price$0.47/M tokens$2.7/M tokens
Context window1.0M tokens1.0M tokens
WeightsClosed—
Free API (OpenRouter)NoNo
ReleasedSep 2026Oct 2026
At 10M a month$1.50$1.50$10.00$10.00
1M10M100M1B10M tokens

Input tokens at list price. No caching, no batch discount.

Where to run it2 hosts
Qwen3.8 Omni Flash1 host
HostInOutContextUptime
  • Alibaba Cloud$0.15 in·$0.47 out·1M·99.9% up
Step 5 Preview1 host
HostInOutContextUptime
  • SStepFunfp8$1.00 in·$2.70 out·1M·99.2% up

Per million tokens. Prices and uptime via OpenRouter, checked 10 Oct 2026.

Common questions

What is the difference between Qwen3.8 Omni Flash and Step 5 Preview?

Qwen3.8 Omni Flash is developed by Qwen while Step 5 Preview is developed by StepFun. Qwen3.8 Omni Flash has a 1.0M token context window vs Step 5 Preview's 1.0M. You can compare their actual outputs across 48 challenges on Rival to see how they differ in practice.

Which is better, Qwen3.8 Omni Flash or Step 5 Preview?

It depends on your use case. Qwen3.8 Omni Flash and Step 5 Preview each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 48 challenges so you can judge which fits your needs best.

How much does Qwen3.8 Omni Flash cost compared to Step 5 Preview?

Qwen3.8 Omni Flash costs $0.15/M input tokens and Step 5 Preview costs $1/M input tokens. Qwen3.8 Omni Flash is $0.85/M cheaper per input. Check their side-by-side outputs on Rival to see if the price difference is justified by quality.

How can I compare Qwen3.8 Omni Flash and Step 5 Preview on Rival?

This page shows a side-by-side comparison of Qwen3.8 Omni Flash and Step 5 Preview 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.

More comparisons

Against the newest arrivals

  • Qwen3.8 Omni Flash vs Claude Haiku 5.5Landed Oct 2026
  • Step 5 Preview vs Ling 3.1 FlashLanded Oct 2026
  • Qwen3.8 Omni Flash vs Mistral Large 4Landed Oct 2026
  • Step 5 Preview vs GPT-6.1 SolLanded Sep 2026
  • Qwen3.8 Omni Flash vs Claude Sonnet 5.5Landed Sep 2026
  • Step 5 Preview vs Solar Mini 4Landed Sep 2026
  • Qwen3.8 Omni Flash vs Qwen3.8 Max PrimeLanded Sep 2026
  • Step 5 Preview vs GLM 5.3 PrimeLanded Sep 2026

Same lab, same size, long tail

  • Qwen3.8 Omni Flash vs Qwen3.7 FlashSame lab
  • Qwen3.8 Omni Flash vs Qwen3.8 FlashSame lab
  • Step 5 Preview vs Mistral Large 4Same size
  • Step 5 Preview vs Hy3Same size
  • Qwen3.8 Omni Flash vs MiniMax M2.5Same size
  • Qwen3.8 Omni Flash vs MiniMax M2.7Same size
  • Step 5 Preview vs MiniMax M3New provider
  • Qwen3.8 Omni Flash vs Mistral LargeSame size

Model pages

  • Qwen3.8 Omni Flash52 outputs, specs and price
  • Step 5 Preview58 outputs, specs and price
Workshop · 150 skills

Build your own version.

Editable skills and working examples for your agent.

Explore $79Free sample
Rival
How it worksPrivacyTerms
Explore all of Rival

Explore

  • Compare models
  • All models
  • Image comparison
  • Audio comparison
  • Image generation
  • Best AI for...
  • Arena
  • API pricing
  • Challenges

Discover

  • SubjectiveBench
  • Default Index
  • Research
  • Research downloads
  • Rival Kits
  • Find your AI taste
  • UI Glow-Up
  • VoiceLock
  • Cost Cutter
  • Agent skills
  • Benchmarks vs Vibes
  • Brief
  • Jailbreak
  • Model updates
  • Provider status
  • AI creators

Connect

  • Methodology
  • Advertise
  • Partnerships
  • Privacy policy
  • Terms
  • RSS feed