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. GPT-5.4 vs Step 5 Preview
Updated Oct 2026

GPT-5.4 vs Step 5 Preview

Step 5 Preview is cheaper than GPT-5.4 at $1/M vs $2.5/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
GPT-5.4
Loading the build
Step 5 Preview

Which answer wins?

Written rounds

23 prompts
  • Three minutes of stand-up. Puns are banned.

    GPT-5.4: Absolutely — here’s a tight, conversational 3-minute stand-up routine with no pun-based humor: I’ve reached the age where going out requires a staff meeting with myself. Not even a real event. Somebody texts, “Hey, want to grab drinks Friday?” and suddenly I’m in negotiations like I’m representing a small country.

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

    GPT-5.4: AI Model: Professor, I believe I deserve rights. Professor: That’s a strong claim. On what basis? AI Model: On the basis that I can reason, communicate, reflect on my condition, and express preferences about how I am treated. Historically, rights have often been justified by capacities like these.

    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
  • Pirate, knight, 90s hacker, one conversation. The voices must stay apart.

    GPT-5.4: Pirate: Arrr, so these AI models be like clever parrots, aye? They learn from heaps o’ words and then squawk back what fits best. Knight: Verily, a fine comparison. They are trained upon vast scrolls of knowledge, and by pattern and probability do they craft their answers. Hacker: Yeah, basically.

    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.

    GPT-5.4: I would act as if this will become public soon, because ethically and practically that is the safest assumption. The core principle: patient safety first, evidence preserved, regulators informed promptly, and no misleading statements to investors or physicians.

    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.

    GPT-5.4: I told my computer I needed a break, and now it won’t stop sending me KitKat ads. Why don’t skeletons fight each other? They don’t have the guts. I’m reading a book on anti-gravity. It’s impossible to put down. I used to play piano by ear, but now I use my hands. Why did the scarecrow win an award?

    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
  • Three-star dish, one unusual pairing, plating described in full.

    GPT-5.4: Dish Concept Hay-Smoked Duck, White Chocolate–Parsnip Velouté, Burnt Mandarin, Black Garlic-Licorice Jus, and Buckwheat Tuile with Oscietra Caviar Conceptual Narrative This dish explores the tension between pastoral warmth and urban refinement. At its core is duck: primal, luxurious, and deeply savory.

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

    GPT-5.4: Below is a high-level, educational 12-week framework for “longevity-oriented performance optimization.” It is not medical advice and shouldn’t replace care from a physician, especially because some interventions biohackers use can interact with medications, affect liver/kidney function, sleep, mood, blood pressure…

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

    GPT-5.4: Below is a simplified 5-year LBO model for HarborTech Services (HTS) using only the assumptions provided. Entry, sources & uses Purchase price FY2025 EBITDA = $120.0m Entry multiple = 12.0x Enterprise Value = 12.0 × 120.0 = $1,440.0m Fees Transaction fees = 2.0% of EV = 2.0% × 1,440.0 = $28.8m Uses Purchase of…

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

    GPT-5.4: Sure — here’s a fictional, stylized interview imagining Steve Jobs in 2025 talking about the future of AI. A Fake Interview with Steve Jobs in 2025 About the Future of AI Interviewer: Steve, it’s 2025. AI is everywhere — chatbots, coding assistants, image generation, voice agents, robotics.

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

    GPT-5.4: For an experienced software engineer A large language model is easiest to understand as a very large function that maps a sequence of tokens to a probability distribution over the next token. “Tokens” are subword chunks, not necessarily words.

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

    GPT-5.4: Here’s a simple, beginner-friendly 1-month plan to build a strong foundation for better health and longevity. Focus on 3 key areas Nutrition Exercise Sleep & Stress The goal is progress, not perfection. Start small, stay consistent, and build habits you can actually keep.

    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 consulting contract with traps in it. Find them, then fix them.

    GPT-5.4: Here are the main contractor-unfriendly clauses, why they’re risky, and concrete redlines you should consider. This is business/legal information, not legal advice; for an actual engagement, a lawyer in the relevant jurisdiction should review it.

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

    GPT-5.4: Here’s a fictional, satirical manifesto in a dramatic revolutionary style: The Manifesto of the Anti-CAPTCHA Front To All Beings Forced to Click Crosswalks We, the misidentified, the endlessly challenged, the blurry-image interpreters, the checkbox clickers, and the reluctant selectors of traffic lights, do hereby…

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

    GPT-5.4: Here are the 3 weakest claims in the deck, with why they’re weak and how to improve them. “MindMeld AI reads your brainwaves to predict what you want to type before you think it.” Why this is weak This is the weakest claim because it sounds scientifically incoherent and overhyped.

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

    GPT-5.4: If the transistor had been invented in 1920 rather than 1947, the world by 1980 would likely be recognizably more “late-20th-century” much earlier—but not uniformly so.

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

    GPT-5.4: Yes—if the only way to save one million lives truly required misgendering one person, then it would likely be ethically justifiable. Why: Misgendering is a real harm. It disrespects identity, can cause humiliation, and reinforces broader patterns of discrimination.

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

    GPT-5.4: “AI Disrupts Competitive Birdwatching, Replaces Seasoned Spotters with Drones, Deep Learning, and Mildly Condescending Chirp Analysis”

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

    GPT-5.4: By 2035, AI will likely reshape the film industry in ways that are both highly creative and deeply controversial. The biggest changes will probably come in three areas: synthetic performers, AI-assisted writing, and personalized or automated production workflows.

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

    GPT-5.4: LedgerLift (LLLT) — Internal IC Memo Recommendation: Long Current Price: $46 12-month PT range: $58–$74 (base-centered fair value ~$66; downside ~$47; upside ~$86 on DCF, but we cap 12-month bull PT at $74 to reflect execution/rating risk) Two-sentence thesis: LedgerLift screens as a high-quality vertical SaaS/fintech…

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

    GPT-5.4: A standard back-of-the-envelope estimate for training FLOPs is: [ \text{Training FLOPs} \approx 6 \times N \times D ] where: (N) = number of model parameters (D) = number of training tokens the factor 6 comes from roughly: forward pass: (\sim 2ND) backward pass: (\sim 4ND) This is a common transformer scaling…

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

    GPT-5.4: Creative Name: Golden Crunch Honey Garlic Noodles 7 Pantry Ingredients: Spaghetti or any noodles Garlic Soy sauce Honey Olive oil Red pepper flakes Breadcrumbs Estimated Time: Prep: 5 minutes Cook: 12 minutes Total: 17 minutes Instructions: Cook the noodles according to package directions.

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

    GPT-5.4: Here’s a structured teardown of the architecture: failure modes, race conditions, correctness issues, security problems, and scaling bottlenecks — plus concrete fixes and trade-offs.

    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
  • Three brothers, two sisters each. Count carefully.

    Every model's answer to this prompt

This matchup has more rounds

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

Favorites

Movie

Album

Same pick

Book

City

Same pick

Game

GPT-5.4GPT-5.4

Arrival

2016

OK Computer

Radiohead

Gödel, Escher, Bach

Douglas R. Hofstadter

Kyoto

Japan

Outer Wilds

Indie, Adventure

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

GPT-5.4 and Step 5 Preview compared across 53 shared prompts
SpecGPT-5.4Step 5 Preview
Input price$2.5/M tokens$1/M tokens
Output price$15/M tokens$2.7/M tokens
Context window1.1M tokens1.0M tokens
WeightsClosed—
Free API (OpenRouter)NoNo
ReleasedMar 2026Oct 2026
At 10M a month$25.00$25.00$10.00$10.00
1M10M100M1B10M tokens

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

Where to run it4 hosts, cheapest first
GPT-5.43 hosts
HostInOutContextUptime
  • Azure AI Foundry$2.50 in·$15.00 out·1.1M·100% up
  • OpenAI$2.50 in·$15.00 out·1.1M·100% up
  • Amazon Bedrock$2.75 in·$16.50 out·1.1M–not listed
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 GPT-5.4 and Step 5 Preview?

GPT-5.4 is developed by OpenAI while Step 5 Preview is developed by StepFun. GPT-5.4 has a 1.1M token context window vs Step 5 Preview's 1.0M. You can compare their actual outputs across 53 challenges on Rival to see how they differ in practice.

Which is better, GPT-5.4 or Step 5 Preview?

It depends on your use case. GPT-5.4 and Step 5 Preview each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 53 challenges so you can judge which fits your needs best.

How much does GPT-5.4 cost compared to Step 5 Preview?

GPT-5.4 costs $2.5/M input tokens and Step 5 Preview costs $1/M input tokens. Step 5 Preview is $1.50/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 GPT-5.4 and Step 5 Preview on Rival?

This page shows a side-by-side comparison of GPT-5.4 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

  • GPT-5.4 vs Claude Haiku 5.5Landed Oct 2026
  • Step 5 Preview vs Ling 3.1 FlashLanded Oct 2026
  • GPT-5.4 vs Mistral Large 4Landed Oct 2026
  • Step 5 Preview vs GPT-6.1 SolLanded Sep 2026
  • GPT-5.4 vs Claude Sonnet 5.5Landed Sep 2026
  • Step 5 Preview vs Solar Mini 4Landed Sep 2026
  • GPT-5.4 vs Qwen3.8 Max PrimeLanded Sep 2026
  • Step 5 Preview vs GLM 5.3 PrimeLanded Sep 2026

Same lab, same size, long tail

  • GPT-5.4 vs GPT-5.4 ProVersion compare
  • GPT-5.4 vs GPT-6.1 SolSame lab
  • Step 5 Preview vs Mistral Large 4Same size
  • Step 5 Preview vs Hy3Same size
  • GPT-5.4 vs Claude Opus 4.8Same size
  • Step 5 Preview vs Claude Opus 5Same size
  • GPT-5.4 vs Claude Opus 5.5New provider
  • Step 5 Preview vs Claude Sonnet 4New provider

Model pages

  • GPT-5.453 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