GPT-5 Mini
VS
Grok 4.20
GPT-5 Mini vs Grok 4.20: which wins at real work?
22 task areas · same graded test runs · rank comparison only, so 0–100 and Elo collections never mix raw scores.
GPT-5 Mini wins 18 of 22 task areas we tested; Grok 4.20 takes 4. GPT-5 Mini costs 1.7× less per token ($2.25 vs $3.75 per 1M).
GPT-5 Mini costs 1.7× less per token ($2.25 vs $3.75 per 1M).
Task by task
| Task area | GPT-5 Mini | Grok 4.20 | Winner |
|---|---|---|---|
| Legal & HR |
#10
/ 110
Excellent
|
#87
/ 110
Strong
|
GPT-5 Mini |
| Data & Analytics |
#27
/ 110
Excellent
|
#102
/ 110
Excellent
|
GPT-5 Mini |
| Chef / Home Cooking |
#45
/ 126
Usable
|
#100
/ 126
Needs editing
|
GPT-5 Mini |
| Product & Project Management |
#1
/ 110
Excellent
|
#56
/ 110
Strong
|
GPT-5 Mini |
| Coding |
#37
/ 115
Strong
|
#86
/ 115
Usable
|
GPT-5 Mini |
| Content & Brand |
#69
/ 124
Strong
|
#118
/ 124
Needs editing
|
GPT-5 Mini |
| Translation & Localization |
#52
/ 110
Excellent
|
#98
/ 110
Strong
|
GPT-5 Mini |
| AI Strategy |
#100
/ 126
Usable
|
#59
/ 126
Strong
|
Grok 4.20 |
| Customer Support |
#104
/ 113
Usable
|
#68
/ 113
Usable
|
Grok 4.20 |
| Research & Competitive Analysis |
#21
/ 110
Excellent
|
#52
/ 110
Usable
|
GPT-5 Mini |
| Structured Output |
#17
/ 113
Excellent
|
#48
/ 113
Excellent
|
GPT-5 Mini |
| Knowledge & Docs |
#41
/ 110
Strong
|
#70
/ 110
Usable
|
GPT-5 Mini |
| Creative & Comedy | #70 / 110 | #96 / 110 | GPT-5 Mini |
| Executive Assistant |
#108
/ 112
Usable
|
#83
/ 112
Usable
|
Grok 4.20 |
| RAG, Safety & Grounding |
#78
/ 113
Excellent
|
#96
/ 113
Strong
|
GPT-5 Mini |
| Sales |
#102
/ 110
Needs editing
|
#85
/ 110
Usable
|
Grok 4.20 |
| Landing Pages |
#50
/ 72
Usable
|
#66
/ 72
Usable
|
GPT-5 Mini |
| Investor & Pitch |
#50
/ 66
Usable
|
#59
/ 66
Usable
|
GPT-5 Mini |
| Presentations & Decks |
#39
/ 110
Excellent
|
#46
/ 110
Excellent
|
GPT-5 Mini |
| Training & Education |
#81
/ 110
Strong
|
#88
/ 110
Strong
|
GPT-5 Mini |
| Summarization & Meeting Notes |
#81
/ 110
Excellent
|
#85
/ 110
Excellent
|
GPT-5 Mini |
| Frontend & Landing Pages |
#12
/ 109
Needs editing
|
#15
/ 109
Needs editing
|
GPT-5 Mini |
Rank = position among every model config we tested in that task area (lower is better). Sorted by biggest gap first.
Frequently asked
Is GPT-5 Mini better than Grok 4.20?
Across 22 task areas we benchmarked, GPT-5 Mini ranks higher in 18 and Grok 4.20 in 4.
Which is cheaper, GPT-5 Mini or Grok 4.20?
GPT-5 Mini costs 1.7× less per token ($2.25 vs $3.75 per 1M).
What is GPT-5 Mini better at?
GPT-5 Mini out-ranks Grok 4.20 at Legal & HR, Data & Analytics, Chef / Home Cooking.
What is Grok 4.20 better at?
Grok 4.20 out-ranks GPT-5 Mini at AI Strategy, Customer Support, Executive Assistant.
More comparisons
This page is Spring Prompt, running in public
We just did this for every model. Do it for your prompt.
The rankings above come from running real tasks through real models and scoring every output. Spring Prompt is that same engine — pointed at your prompt, your test cases, and your definition of good.
- Generate test cases from your prompt — no eval set required to start.
- Compare models side by side with quality, cost and latency in one matrix.
- Optimise the winner until the scores say it's ready to ship.
Prompt × model results
12 test cases · 3 evals