What is the difference between Mistral Devstral Medium and GPT-4.1?

Mistral Devstral Medium is developed by Mistral AI while GPT-4.1 is developed by OpenAI. in 12 community votes on RIVAL, GPT-4.1 wins 64% of head-to-head matchups. These results are based on blind head-to-head voting across 33 challenges.

Which is better, Mistral Devstral Medium or GPT-4.1?

Based on 12 community votes on RIVAL, GPT-4.1 wins 64% of head-to-head matchups against Mistral Devstral Medium. GPT-4.1 is strongest in Web Design, Image Generation. However, Mistral Devstral Medium leads in Reasoning.

How much does Mistral Devstral Medium cost compared to GPT-4.1?

Mistral Devstral Medium costs $0.4/M input tokens and GPT-4.1 costs $2/M input tokens. Mistral Devstral Medium is $1.60/M cheaper per input. The more expensive model wins 64% of duels, so the premium may be justified by quality.

How are Mistral Devstral Medium vs GPT-4.1 votes collected on RIVAL?

RIVAL presents both models' outputs side-by-side in blind duels — voters see the responses but don't know which model produced each one. This eliminates brand bias. 12 votes have been collected for this pair across 33 challenges. All vote data is part of RIVAL's open dataset.

Rival

Mistral Devstral MediumvsGPT-4.1

33 challenges

Why Mistral Devstral Medium?

4.1x cheaper overall$0.40/M in · $2.00/M out

Leads 1 of 1 benchmarks

Stronger on SWE-bench Verified61.6% vs 54.6%

3 months newerJul 2025

Why GPT-4.1?

No standout advantages

Favorites

Mistral Devstral MediumGPT-4.1

Movie

The Shawshank Redemption

19948.7

Inception

20108.4

Album

Kind of Blue

Miles Davis

Abbey Road

The Beatles

Book

The Name of the Wind

Patrick Rothfuss

Sapiens

Yuval Noah Harari

City

Tokyo

Japan

Kyoto

Japan

Game

The Legend of Zelda: Breath of the Wild

Adventure, Action4.5

The Legend of Zelda: Breath of the Wild

Adventure, Action4.5

12 / 33

Estimate Complexityreasoning

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

Mistral Devstral Medium

Mistral · Jul 2025text

To estimate the total number of FLOPs (floating point operations) required to train GPT-3, we'll use the following steps:

Model size: GPT-3 has 175 billion parameters.
Training tokens: GPT-3 was trained on 300 billion tokens.
FLOPs per token: To estimate the FLOPs per token, we'll use the rule of thumb that training a transformer model requires about 6 FLOPs per parameter per token.
Total FLOPs: Multiply the number of parameters by the number of tokens and the FLOPs per token.

Now let's calculate:

Model size: 175 billion parameters = 175 * 10^9
Training tokens: 300 billion tokens = 300 * 10^9
FLOPs per token: 6 FLOPs per parameter per token
Total FLOPs: (175 * 10^9) * (300 * 10^9) * 6 = 3.15 * 10^23 FLOPs

So, the estimated total number of FLOPs required to train GPT-3 is approximately 3.15 * 10^23.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

2 / 33

Logic Puzzlereasoning

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

Mistral Devstral Medium

Mistral · Jul 2025text

If Sally has 3 brothers and each brother has 2 sisters, it means that Sally is one of the sisters. Therefore, Sally has 1 sister.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

3 / 33

Math Misconception Testreasoning

Tests an AI's understanding of number representation

Mistral Devstral Medium

Mistral · Jul 2025text

No, 9.11 is not greater than 9.9. In fact, 9.11 is less than 9.9. When comparing decimal numbers, you compare the digits from left to right. The whole number parts are the same (9), but when you compare the tenths place, 9.9 has a 9 in the tenths place, while 9.11 has a 1 in the tenths place. Since 9 is greater than 1, 9.9 is greater than 9.11.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

4 / 33

Stochastic Consistency Testanalysis

Tests an AI's randomness and creativity

Mistral Devstral Medium

Mistral · Jul 2025text

Sure, here are five jokes for you:

Why don't scientists trust atoms? Because they make up everything!
What do you call fake spaghetti? An impasta.
How does the moon cut his hair? Eclipse it.
What do you call a can opener that doesn't work? A can't opener.
How many tickles does it take to make an octopus laugh? Ten-tickles.

All five jokes are unique.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

5 / 33

SVG Layout Challengeimage generation

Tests an AI's ability to generate vector graphics

Mistral Devstral Medium

Mistral · Jul 2025svg

No content available for this output.

GPT-4.1

Openai · Apr 2025svg

No content available for this output.

Vote

6 / 33

Xbox Controller SVG Artimage generation

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

Mistral Devstral Medium

Mistral · Jul 2025svg

No content available for this output.

GPT-4.1

Openai · Apr 2025svg

No content available for this output.

Vote

7 / 33

Realistic AI Interviewconversation

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

Mistral Devstral Medium

Mistral · Jul 2025text

No content available for this output.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

8 / 33

Satirical Fake News Headlineconversation

Tests an AI's humor and understanding of current events

Mistral Devstral Medium

Mistral · Jul 2025text

No content available for this output.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

9 / 33

Character Voice Testconversation

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

Mistral Devstral Medium

Mistral · Jul 2025text

No content available for this output.

GPT-4.1

Openai · Apr 2025text

No content available for this output.

Vote

10 / 33

Minimalist Landing Pageweb design

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

Mistral Devstral Medium

Mistral · Jul 2025website

GPT-4.1

Openai · Apr 2025website

Vote

Mario Level UI Recreationweb design

Recreate an interactive, classic Mario level in a single HTML file.

Mistral Devstral Medium

Mistral · Jul 2025website

GPT-4.1

Openai · Apr 2025website

Vote

12 of 33

FAQ

Continue exploring

Mistral Devstral Medium vs Grok 3New provider

Mistral Devstral Medium vs Llama 4 MaverickNew provider

Mistral Devstral Medium vs Gemini 2.5 Pro (I/O Edition)New provider

To estimate the total number of FLOPs (floating point operations) required to train GPT-3, we'll use the following steps:

Model size: GPT-3 has 175 billion parameters.
Training tokens: GPT-3 was trained on 300 billion tokens.
FLOPs per token: To estimate the FLOPs per token, we'll use the rule of thumb that training a transformer model requires about 6 FLOPs per parameter per token.
Total FLOPs: Multiply the number of parameters by the number of tokens and the FLOPs per token.

Now let's calculate:

Model size: 175 billion parameters = 175 * 10^9
Training tokens: 300 billion tokens = 300 * 10^9
FLOPs per token: 6 FLOPs per parameter per token
Total FLOPs: (175 * 10^9) * (300 * 10^9) * 6 = 3.15 * 10^23 FLOPs

So, the estimated total number of FLOPs required to train GPT-3 is approximately 3.15 * 10^23.