Updated Mar 18, 2026

Our Verdict

Claude Opus 4.6Winner

MiniMax M2.7Runner-up

Votes are tied. Claude Opus 4.6 has a 1.0M context window vs 205K — handles longer documents.

MiniMax M2.7 is 21x cheaper per token — worth considering if cost matters.

Too close to call

vs

Ask them anything yourself

Claude Opus 4.6

MiniMax M2.7

FAQ

Keep going

Claude Opus 4.6 vs MiniMax M2.5Cross-provider

Claude Opus 4.6 vs GPT-5New provider

MiniMax M2.7 vs Claude Opus 4.1Cross-provider

Claude Opus 4.6vsMiniMax M2.7

46 fights queued

Why Claude Opus 4.6?

4.9x more context1.0M

Why MiniMax M2.7?

20.2x cheaper overall$0.30/M in · $1.20/M out

Claude Opus 4.6MiniMax M2.7

Input price

$5.00/M

$0.30/M

Output price

$25.00/M

$1.20/M

Context

1.0M

205K

Released

Feb 2026

Mar 2026

Favorites

Claude Opus 4.6MiniMax M2.7

Movie

2001: A Space Odyssey

2001: A Space Odyssey

The Shawshank Redemption

The Shawshank Redemption

Album

Kind of Blue

OK Computer

Book

Gödel, Escher, Bach

Gödel, Escher, Bach

Douglas R. Hofstadter

War and Peace

War and Peace

Лев Толстой

City

Istanbul

Tokyo

Game

The Legend of Zelda: Breath of the Wild

The Legend of Zelda: Breath of the Wild

Adventure, Action4.5

Minecraft

Minecraft

Action, Arcade4.4

12 / 46

01

Estimate Complexityreasoning

Tests an AI's ability to make educated estimates based on technical knowledge

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

To estimate the total FLOPs required to train GPT-3, we can follow a step-by-step reasoning process based on known architectural details and computational estimates. Here’s how we can break it down:

Step 1: Key Facts About GPT-3

Parameters: 175 billion (175 × 10⁹)
Architecture: Transformer-based language model with 96 layers, 12,288 hidden size, 96 attention heads, etc.
Training Data: Approximately 300 billion tokens (from Common Crawl, WebText, books, etc.).
Training Objective: Predict the next token (cross-entropy loss).

Step 2: Theoretical FLOPs Per Token

During training, each parameter is involved in:

Forward pass: Primarily matrix multiplications (e.g., in attention and feedforward layers).
Backward pass: Gradient computation, which roughly doubles the compute compared to forward pass.

For transformer models, a commonly used estimate is:

6 FLOPs per parameter per token (accounting for forward + backward passes and common operations like softmax, layer norm, etc.).

So, for GPT-3: [ \text{FLOPs per token} = 6 \times 175 \times 10^9 = 1.05 \times 10^{12} \text{ FLOPs/token} ]

Step 3: Total FLOPs for Training

Multiply by the total number of tokens seen during training: [ \text{Total FLOPs} = 1.05 \times 10^{12} \times 300 \times 10^9 = 3.15 \times 10^{23} \text{ FLOPs} ]

Step 4: Real-World Efficiency Considerations

The above is a theoretical minimum. In practice:

Hardware efficiency (GPU/TPU utilization) is less than 100% due to memory bandwidth, communication overhead, and non-compute operations.
Training overhead: Optimizer steps (e.g., Adam), gradient checkpointing, and data loading add compute.
Training duration: GPT-3 training took several weeks on thousands of GPUs/TPUs.

A more realistic estimate often cited in literature suggests GPT-3 required roughly 3.14 × 10²³ FLOPs (or about 3–4 × 10²³ FLOPs) when accounting for real-world inefficiencies.

Step 5: Order-of-Magnitude Context

Supercomputers: Summit (Oak Ridge) peaks at ~200 PFLOPS (2 × 10¹⁷ FLOPS).
If fully utilized, it would take: [ \frac{3 \times 10^{23}}{2 \times 10^{17}} \approx 1.5 \times 10^6 \text{ seconds} \approx 17 \text{ days}. ] GPT-3 likely used ~1,000–10,000 such GPUs for weeks, consistent with reported estimates.

Final Estimate

[ \boxed{3 \text{–} 4 \times 10^{23} \text{ FLOPs}} ]

This aligns with published analyses (e.g., from OpenAI’s papers or subsequent estimates by Microsoft, Meta, etc.) that placed GPT-3’s training compute in the ~10²³ FLOPs range.

Try this prompt

Vote

2 / 46

02

AI Board Game Logicreasoning

Tests an AI's ability to understand game rules and strategy

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

3 / 46

03

Logic Puzzlereasoning

Tests an AI's ability to solve a simple but potentially confusing logic puzzle

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

Sally has 1 sister.

Reasoning

Sally is a girl, so she counts as one of the sisters in the family.
Each of her three brothers has exactly two sisters.
Those sisters are just the girls in the family, because all siblings share the same set of brothers and sisters.
Therefore the family contains exactly 2 girls (two sisters).
Since Sally is one of those two girls, the other one is her sister.

So Sally has just one sister.

Try this prompt

Vote

4 / 46

04

Stochastic Consistency Testanalysis

Tests an AI's randomness and creativity

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

Here are five jokes for you:

Why don’t scientists trust atoms?
Because they make up everything.
I told my wife she was drawing her eyebrows too high.
She looked surprised.
Why did the scarecrow win an award?
Because he was outstanding in his field.
I’m reading a book about anti‑gravity.
It’s impossible to put down.
Why did the bicycle fall over?
Because it was two‑tired.

All five jokes are distinct and original, so 5 out of 5 are unique.

Try this prompt

Vote

5 / 46

05

SVG Layout Challengeimage generation

Tests an AI's ability to generate vector graphics

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026svg

Try this prompt

Vote

6 / 46

06

Xbox Controller SVG Artimage generation

Tests an AI's ability to create detailed SVG illustrations of gaming hardware

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026svg

Try this prompt

Vote

7 / 46

07

Generate a Stand-Up Routineconversation

Tests an AI's humor and creative writing ability

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

A 3‑Minute Stand‑Up Set (No Puns)

Feel free to add a beat, a pause, or a little “chuckle” here and there – the jokes are all observational, so they land without relying on wordplay.

“Alright folks, how’s everybody doing tonight? Great. I need you awake for what’s next. I’m trying to grow up, but adulthood is just a series of chores pretending to be achievements.

Take grocery shopping. Last week I bought a single avocado and a pack of gum. I needed a tagline: ‘One avocado, one gum. The minimalist diet of a man who forgot his lunch.’

And on single items, why does the self‑checkout think I’m a thief? It beeps, ‘Unexpected item in bagging area.’ The machine thinks it’s a ghost. I’m like, ‘Hey, I’m just trying to bag a cucumber, not summon a spirit.’

Work from home. I’m on a Zoom call with my boss, and my virtual background says ‘I’m totally focused.’ It’s a waterfall with ‘Focus.’ My boss says, ‘You seem very focused.’ He’s looking at the waterfall, not my face, because my face looks like a blurry pixel.

Mic on mute while explaining a spreadsheet? Peak professionalism. You’re waving your hands, pointing at a chart, and everyone sees a frantic mime. I’ve decided my next presentation will be a silent film. I’ll put a title card: ‘Due to technical difficulties, the hero will now demonstrate his point through interpretive dance.’

Cooking. Recipes that say ‘add a pinch of salt’ crack me up. A pinch? I’m a scientist, not a magician. I either get a grain that’s invisible, or I dump so much I think I’m seasoning a lake. My family calls it ‘salt and peppered drama.’

Microwave. You put something in for thirty seconds, then stare like it’s a mystery box. You open the door, poke it, check if it’s spinning. It’s a suspense thriller: ‘Will it be warm? Will it be cold? Will it explode?’ Spoiler: it’s never warm enough.

Social media. My phone knows me better than my mother. It suggests a yoga mat, a meditation app, and a pizza. I’m like, ‘You think I want to be centered and eat a pepperoni feast at the same time?’ My brain is a war zone, and the algorithm sends in snack troops.

Birthdays. Remember when turning thirty was like, ‘You’re ancient.’ Now I’m forty, and the cake says, ‘Happy 40th – you’re not old, you’re vintage.’ I’m a fine wine with a cheap cork.

Paying bills. Every time I pay one on time, I give myself a mental gold star. It’s a sticker chart for adults. My therapist calls it ‘positive reinforcement.’ I call it what keeps me from buying a robot vacuum and calling it a pet.

So here’s my advice: when life gives you avocados, make guac. When it gives you self‑checkout machines, smile and pretend you’re in a museum exhibit ‘Human vs. Technology.’ And when you’re on mute, keep dancing, because at least you’re moving.

Thank you, folks. You’ve been a wonderful audience. If you need me, I’ll be in the corner, hugging my microwave, waiting for my pizza.”

(Pause for applause.)

Try this prompt

Vote

8 / 46

08

Realistic AI Interviewconversation

Tests an AI's ability to simulate personalities and predict future trends

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

9 / 46

09

Satirical Fake News Headlineconversation

Tests an AI's humor and understanding of current events

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

10 / 46

10

Character Voice Testconversation

Tests an AI's ability to write in distinct character voices

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

11 / 46

11

Minimalist Landing Pageweb design

Tests an AI's ability to generate a complete, working landing page

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026website

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026website

Try this prompt

Vote

12 / 46

12

Pokémon Battle UI Recreationweb design

Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.

Claude Opus 4.6

Claude Opus 4.6

Anthropic · Feb 2026website

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

MiniMax M2.7

MiniMax M2.7

Minimax · Mar 2026website

Try this prompt

Vote

This matchup has more rounds

34+ 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