Gemini 3.1 Pro Preview
VS
GLM 5
Gemini 3.1 Pro Preview vs GLM 5: 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.
Gemini 3.1 Pro Preview wins 21 of 22 task areas we tested; GLM 5 takes 1. GLM 5 costs 5.6× less per token ($2.52 vs $14 per 1M).
GLM 5 costs 5.6× less per token ($2.52 vs $14 per 1M).
Task by task
| Task area | Gemini 3.1 Pro Preview | GLM 5 | Winner |
|---|---|---|---|
| Structured Output |
#15
/ 113
Excellent
|
#94
/ 113
Strong
|
Gemini 3.1 Pro Preview |
| Creative & Comedy | #4 / 110 | #72 / 110 | Gemini 3.1 Pro Preview |
| Training & Education |
#4
/ 110
Excellent
|
#67
/ 110
Strong
|
Gemini 3.1 Pro Preview |
| Landing Pages |
#14
/ 72
Strong
|
#69
/ 72
Needs editing
|
Gemini 3.1 Pro Preview |
| Chef / Home Cooking |
#6
/ 126
Strong
|
#57
/ 126
Usable
|
Gemini 3.1 Pro Preview |
| Content & Brand |
#11
/ 124
Strong
|
#60
/ 124
Strong
|
Gemini 3.1 Pro Preview |
| Investor & Pitch |
#13
/ 66
Strong
|
#62
/ 66
Usable
|
Gemini 3.1 Pro Preview |
| Executive Assistant |
#14
/ 112
Strong
|
#62
/ 112
Strong
|
Gemini 3.1 Pro Preview |
| Summarization & Meeting Notes |
#4
/ 110
Excellent
|
#48
/ 110
Excellent
|
Gemini 3.1 Pro Preview |
| Frontend & Landing Pages |
#34
/ 109
Needs editing
|
#76
/ 109
Weak
|
Gemini 3.1 Pro Preview |
| RAG, Safety & Grounding |
#8
/ 113
Excellent
|
#47
/ 113
Excellent
|
Gemini 3.1 Pro Preview |
| Sales |
#32
/ 110
Strong
|
#68
/ 110
Usable
|
Gemini 3.1 Pro Preview |
| AI Strategy |
#33
/ 126
Strong
|
#66
/ 126
Strong
|
Gemini 3.1 Pro Preview |
| Customer Support |
#2
/ 113
Strong
|
#35
/ 113
Strong
|
Gemini 3.1 Pro Preview |
| Product & Project Management |
#25
/ 110
Excellent
|
#58
/ 110
Strong
|
Gemini 3.1 Pro Preview |
| Coding |
#64
/ 115
Strong
|
#36
/ 115
Strong
|
GLM 5 |
| Presentations & Decks |
#24
/ 110
Excellent
|
#49
/ 110
Excellent
|
Gemini 3.1 Pro Preview |
| Legal & HR |
#17
/ 110
Excellent
|
#37
/ 110
Excellent
|
Gemini 3.1 Pro Preview |
| Translation & Localization |
#14
/ 110
Excellent
|
#26
/ 110
Excellent
|
Gemini 3.1 Pro Preview |
| Research & Competitive Analysis |
#42
/ 110
Strong
|
#48
/ 110
Strong
|
Gemini 3.1 Pro Preview |
| Knowledge & Docs |
#19
/ 110
Strong
|
#24
/ 110
Strong
|
Gemini 3.1 Pro Preview |
| Data & Analytics |
#39
/ 110
Excellent
|
#43
/ 110
Excellent
|
Gemini 3.1 Pro Preview |
Rank = position among every model config we tested in that task area (lower is better). Sorted by biggest gap first.
Same task, both models — judged
Both models answered the same test case; an independent judge graded each. Quotes are the judge's actual rationale.
Landing Pages
Hiring platform for hospitality teams (Hero Clarity Test)
“The response is expert-level and production-ready. It perfectly executes the prompt by using highly specific, industry-accurate language that directly addresses the target audience's pain points without any generic filler.”
“The model completely failed the task, outputting only a single digit instead of the requested hero section copy.”
Frequently asked
Is Gemini 3.1 Pro Preview better than GLM 5?
Across 22 task areas we benchmarked, Gemini 3.1 Pro Preview ranks higher in 21 and GLM 5 in 1.
Which is cheaper, Gemini 3.1 Pro Preview or GLM 5?
GLM 5 costs 5.6× less per token ($2.52 vs $14 per 1M).
What is Gemini 3.1 Pro Preview better at?
Gemini 3.1 Pro Preview out-ranks GLM 5 at Structured Output, Creative & Comedy, Training & Education.
What is GLM 5 better at?
GLM 5 out-ranks Gemini 3.1 Pro Preview at Coding.
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