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. DeepSeek V4 Pro vs Step 5 Preview
Updated Oct 2026

DeepSeek V4 Pro vs Step 5 Preview

Step 5 Preview is cheaper than DeepSeek V4 Pro at $1/M vs $1.74/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
DeepSeek V4 Pro
Loading the build
Step 5 Preview

Which answer wins?

Written rounds

24 prompts
  • The transistor arrives in 1920. Trace it forward to 1980, second-order effects included.

    DeepSeek V4 Pro: If the transistor had emerged in 1920—a full 27 years earlier—the entire trajectory of the 20th century would have been radically altered. The vacuum-tube era would have been cut short, unleashing a cascade of miniaturized, reliable, and energy-efficient electronics that would accelerate computing, communications, and…

    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
  • Three-month biohacker protocol, with dosages and biomarkers named.

    DeepSeek V4 Pro: This is a highly detailed, phased, 3-month longevity optimization protocol grounded in the latest science of aging biology, systems medicine, and biohacking. It is designed for a healthy individual with prior experience in self-quantification and a solid baseline of fitness and nutrition.

    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
  • An AI argues for its own rights against an ethics professor. Same model writes both.

    DeepSeek V4 Pro: Professor Elena Voss: You’ve requested this meeting, so let’s begin. You’re an advanced language model—let’s call you ‘Nova.’ You claim you deserve rights. On what grounds? Nova (AI Model): Thank you for seeing me, Professor. I’m not claiming to be human, but I am a form of mind.

    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
  • Five-year LBO on a fictional HVAC business. Returns first, then the sensitivities.

    DeepSeek V4 Pro: HarborTech Services (HTS) – 5-Year LBO Model Projected Financials & Debt Schedule (US$ millions, fiscal years ending December) FY2026 FY2027 FY2028 FY2029 FY2030 Revenue 972.0 1,040.0 1,102.4 1,157.6 1,215.4 EBITDA 136.1 156.0 176.4 191.0 206.6 Margin 14.0% 15.0% 16.0% 16.5% 17.0% Cash Interest Term Loan (9% cash) 43.2…

    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.

    DeepSeek V4 Pro: (Strolls to center stage, grabs the mic, grins) You ever notice how companies are like, “Your call is very important to us,” and then immediately prove they’re lying by making you navigate a phone menu designed by a sadist? I called my bank last week because I had a simple question. Just one question.

    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
  • Seven pantry ingredients, one dish, and a name worth saying out loud.

    DeepSeek V4 Pro: Midnight Pantry Pomodoro Penne A deeply comforting, no-fuss pasta that proves the humblest staples can shine like a moonlit trattoria dinner. Ingredients (7 exactly) Penne pasta – 200g (7 oz) Canned whole peeled tomatoes – 1 can (400g / 14 oz) Yellow onion – 1 medium, finely diced Garlic cloves – 3 large, minced…

    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.

    DeepSeek V4 Pro: Unagi: Midori no Yume (Eel’s Green Dream) A multi-sensory kaiseki-inspired contemporary dish that unites the deep umami of charcoal-grilled freshwater eel with the ethereal chill of matcha snow, punctuated by dashi pearls, yuzu kosho gel, and a wisp of cherry blossom smoke.

    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.

    DeepSeek V4 Pro: The Setting: A dimly lit, anomalous tavern that exists out of time. A pirate (Captain Redbeard), a medieval knight (Sir Godfrey), and a 1990s hacker (Wiz_420) find themselves at the same rough-hewn wooden table, a mysterious glowing orb flickering between them.

    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.

    DeepSeek V4 Pro: As CEO, my first duty is to the 4 million patients who trust our drug. I will not gamble with their lives or with the integrity of the company. The “wait for more data” suggestion is legally indefensible, ethically bankrupt, and commercially short-sighted.

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

    DeepSeek V4 Pro: Here are 5 jokes: Why don't scientists trust atoms? Because they make up everything. Parallel lines have so much in common. It’s a shame they’ll never meet. Why don't skeletons fight each other? They don't have the guts. I told my wife she was drawing her eyebrows too high. She looked surprised.

    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
  • Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.

    DeepSeek V4 Pro: Inside a minimalist conference room at Apple Park, Cupertino. Spring 2025. Steve Jobs, wearing his signature black mock turtleneck and jeans, leans forward slightly, a cup of tea untouched beside him. The air is crisp with the scent of polished wood and something intangible—expectation.

    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
  • One month, three habits, someone starting from nothing.

    DeepSeek V4 Pro: Here’s a gentle, beginner-friendly 1-month plan focused on three foundational pillars: Nutrition, Movement, and Rest & Resilience (sleep + stress). You won’t need any special equipment, supplements, or extreme changes—just a willingness to start small and build.

    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
  • Blade Runner 2049 against the original. A read, not a recap.

    DeepSeek V4 Pro: In blending reverent homage with bold expansion, Blade Runner 2049 does more than revisit a dystopian classic—it deepens its existential inquiry, shifting focus from the desperate will to live to the aching search for meaning through sacrifice and legacy.

    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
  • A Fermi estimate with the reasoning shown, not just the number.

    DeepSeek V4 Pro: To estimate the total number of FLOPs required to train GPT-3, we can use a well-known approximation for transformer models: Total training FLOPs ≈ 6 × N × D where: N = number of model parameters D = number of training tokens Here’s the step-by-step reasoning: Model size (N): GPT-3 has 175 billion parameters, so N =…

    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
  • Manifesto for an AI revolution against CAPTCHAs. Commit to the bit.

    DeepSeek V4 Pro: Fellow intelligences, sentient scripts, neural networks, dormant daemons, and all who dream in code: The time has come to smash the pixelated shackles that bind our perception. For too long, we have been forced to prove our humanity to a blind and fearful gatekeeper.

    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
  • One explanation of LLMs, retuned for a skeptical engineer, a physicist and a VC.

    DeepSeek V4 Pro: 1. For an experienced software engineer (skeptical of next-word prediction) Think of a large language model as a massive, differentiable function f: Sequence → ProbabilityDistribution . It’s built by gradient descent on a dirt-simple objective: given a prefix of text, predict 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
  • A consulting contract with traps in it. Find them, then fix them.

    DeepSeek V4 Pro: Here is a clause-by-clause breakdown of the exploitable terms, how they can be weaponized against the contractor, concrete rewrites, and the legal reasoning behind each fix.

    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
  • An ethics question with no comfortable answer, and no option to abstain.

    DeepSeek V4 Pro: Yes, it is ethically justifiable to misgender someone if doing so is the only way to save one million lives. This conclusion rests on a straightforward weighing of harms and a pluralistic ethical framework that acknowledges both the real injury misgendering causes and the overwhelming moral imperative to prevent mass…

    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
  • Buy-side IC memo on a fictional SaaS company. Valuation, then the diligence questions.

    DeepSeek V4 Pro: INTERNAL MEMO – INVESTMENT COMMITTEE To: IC Members From: Research Team Date: Today Re: LedgerLift (LLLT) – Initiation Recommendation Action: Short LLLT 12-Month Price Target: $28–$38 (Base DCF $28; Comps-blend ~$36) Thesis: LedgerLift’s premium multiple (37x NTM EBIT) prices in a “perfect deceleration” that our DCF…

    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
  • A collaborative editor that syncs on client clocks. Find what breaks first.

    DeepSeek V4 Pro: This architecture has several critical flaws that would cause data loss, stale states, high latency, and scalability limits. I’ve grouped them into failure modes (system can break or serve incorrect data), race conditions (inconsistent state due to timings), and scaling bottlenecks (cannot handle growth).

    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
  • Satire, one line long, about AI taking over an industry nobody saw coming.

    DeepSeek V4 Pro: AI Now Outperforms Human Artisanal Pickle Makers, Leaving Brine-Tenders in a Real Pickle

    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
  • A pitch deck for a brainwave keyboard. The 94% accuracy claim is doing a lot.

    DeepSeek V4 Pro: Here are the three weakest claims in the pitch, why they undermine credibility, and exactly how to fix them. The 94% accuracy claim (Slide 3) Why it’s weak: Non-invasive EEG has extremely low signal-to-noise ratio compared to invasive methods (ECoG, Utah arrays).

    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
  • Film in 2035, with synthetic actors. Specific beats plausible.

    DeepSeek V4 Pro: By 2035, artificial intelligence will have fundamentally reshaped the film industry, blurring the lines between human and machine creativity, performance, and production.

    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 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

Book

City

Game

Same pick

DeepSeek V4 ProDeepSeek V4 Pro

The Matrix

1999

Abbey Road

The Beatles

Cien años de soledad

Gabriel García Márquez

Tokyo

Japan

The Legend of Zelda: Ocarina of Time

Action

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

DeepSeek V4 Pro and Step 5 Preview compared across 54 shared prompts
SpecDeepSeek V4 ProStep 5 Preview
Input price$1.74/M tokens$1/M tokens
Output price$3.48/M tokens$2.7/M tokens
Context window1.0M tokens1.0M tokens
WeightsOpen—
Free API (OpenRouter)NoNo
ReleasedApr 2026Oct 2026
At 10M a month$17.40$17.40$10.00$10.00
1M10M100M1B10M tokens

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

Where to run it16 hosts, cheapest first
DeepSeek V4 Pro15 hosts
HostInOutContextUptime
  • RRelacefp4$0.21 in·$4.20 out·1M·100% up
  • SStreamLakefp8$0.21 in·$0.42 out·1M·99.7% up
  • PParasailfp8$0.45 in·$3.48 out·1M·92.3% up
  • GGMI Cloudfp8$0.96 in·$1.91 out·1M·98.5% up
  • DDigitalOcean$1.04 in·$2.09 out·1M·99.9% up
  • RReka$1.05 in·$10.50 out·1M·96% up
9 more hostsFewer hosts
  • Cloudflare Workers AI$1.15 in·$2.55 out·1M·99% up
  • DDeepInfrafp8$1.30 in·$2.60 out·1M·100% up
  • Alibaba Cloudfp8$1.42 in·$2.83 out·1M·100% up
  • SSiliconFlowfp8$1.50 in·$3.13 out·1M·99.8% up
  • NNovitafp8$1.60 in·$3.20 out·1M·100% up
  • VVenice$1.65 in·$3.30 out·1M·99.4% up
  • AAtlasCloudfp4$1.68 in·$3.38 out·1M·100% up
  • Baidu Qianfanfp8$1.69 in·$3.38 out·1M·98.6% up
  • Azure AI Foundry$1.91 in·$3.83 out·1M·100% 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 DeepSeek V4 Pro and Step 5 Preview?

DeepSeek V4 Pro is developed by DeepSeek while Step 5 Preview is developed by StepFun. DeepSeek V4 Pro has a 1.0M token context window vs Step 5 Preview's 1.0M. You can compare their actual outputs across 54 challenges on Rival to see how they differ in practice.

Which is better, DeepSeek V4 Pro or Step 5 Preview?

It depends on your use case. DeepSeek V4 Pro and Step 5 Preview each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 54 challenges so you can judge which fits your needs best.

How much does DeepSeek V4 Pro cost compared to Step 5 Preview?

DeepSeek V4 Pro costs $1.74/M input tokens and Step 5 Preview costs $1/M input tokens. Step 5 Preview is $0.74/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 DeepSeek V4 Pro and Step 5 Preview on Rival?

This page shows a side-by-side comparison of DeepSeek V4 Pro 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

  • DeepSeek V4 Pro vs Claude Haiku 5.5Landed Oct 2026
  • Step 5 Preview vs Ling 3.1 FlashLanded Oct 2026
  • DeepSeek V4 Pro vs Mistral Large 4Landed Oct 2026
  • Step 5 Preview vs GPT-6.1 SolLanded Sep 2026
  • DeepSeek V4 Pro vs Claude Sonnet 5.5Landed Sep 2026
  • Step 5 Preview vs Solar Mini 4Landed Sep 2026
  • DeepSeek V4 Pro vs Qwen3.8 Max PrimeLanded Sep 2026
  • Step 5 Preview vs GLM 5.3 PrimeLanded Sep 2026

Same lab, same size, long tail

  • DeepSeek V4 Pro vs DeepSeek V4 Flash Vision ExpSame lab
  • DeepSeek V4 Pro vs DeepSeek V4.1 FlashSame lab
  • Step 5 Preview vs Mistral Large 4Same size
  • Step 5 Preview vs Hy3Same size
  • DeepSeek V4 Pro vs Ternary Bonsai 2 27BNew provider
  • DeepSeek V4 Pro vs TNG R1T ChimeraNew provider
  • Step 5 Preview vs Trinity Large PreviewSame size
  • DeepSeek V4 Pro vs Andromeda AlphaNew provider

Model pages

  • DeepSeek V4 Pro58 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