DeepSeek-R1-Distill-Qwen-14B against DeepSeek-R1-Distill-Qwen-32B

Pick two models to see them side by side: Overall score, every category, the questions where one did better, and speed and memory. Each model is shown at its newest tested version. Back to the leaderboard.

The short version

Same model line in two sizes: DeepSeek-R1-Distill-Qwen-14B has 14.8 billion parameters and DeepSeek-R1-Distill-Qwen-32B has 32.8 billion parameters.

DeepSeek-R1-Distill-Qwen-32B scores higher overall: 89.1 against 85.3 out of 100, a gap of 3.8 points.

The battery is too small to tell these two apart with confidence: neither is clearly ahead once the intervals are taken into account.

Across the 9 categories that count toward Overall, DeepSeek-R1-Distill-Qwen-14B is higher in 2, DeepSeek-R1-Distill-Qwen-32B is higher in 6 and 1 is level.

The widest gap is in Long context: 77.8 for DeepSeek-R1-Distill-Qwen-14B against 100.0 for DeepSeek-R1-Distill-Qwen-32B, in favour of DeepSeek-R1-Distill-Qwen-32B.

Of the 259 questions both models have a definite score for, DeepSeek-R1-Distill-Qwen-14B did better on 17, DeepSeek-R1-Distill-Qwen-32B did better on 22, and 220 were level.

On the same backend and Ollama version, with one request at a time, DeepSeek-R1-Distill-Qwen-14B generates 7.1 tokens per second and DeepSeek-R1-Distill-Qwen-32B generates 3.1.

Written from the numbers below. Scores come from one fixed battery run on one laptop; read the methodology before drawing conclusions.

Overall

DeepSeek-R1-Distill-Qwen-32B scores higher: 89.1 against 85.3, a gap of 3.8 points. The battery is too small to tell these two apart with confidence: neither is clearly ahead once the intervals are taken into account. How sure is a score?

Overall (0–100) with its 95% interval, the rank range among all listed ranked models, and the hard-question score.
MeasureDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
Overall85.381.2–89.289.186.0–91.9
Rank range38–10619–85
Hard questionsmean score on the questions that half or fewer of the ranked models (chosen per run, see the methodology) answered correctly——
Size class9–16B16–35B

By category

Programmatic mean (0–100) per category. Higher is the model with the larger mean; a gap that rounds to 0.0 is called level. A dash means not scored, which is different from 0.
CategoryDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32BHigherDifference
Reasoningcounts toward Overall 91.4 97.1 DeepSeek-R1-Distill-Qwen-32B 5.7
Mathcounts toward Overall 88.0 92.0 DeepSeek-R1-Distill-Qwen-32B 4.0
Codingcounts toward Overall 53.0 53.6 DeepSeek-R1-Distill-Qwen-32B 0.6
Instruction followingcounts toward Overall 80.8 75.7 DeepSeek-R1-Distill-Qwen-14B 5.1
Knowledge & honestycounts toward Overall 100.0 100.0 Level 0
Long contextcounts toward Overall 77.8 100.0 DeepSeek-R1-Distill-Qwen-32B 22.2
Multilingualcounts toward Overall 89.2 96.7 DeepSeek-R1-Distill-Qwen-32B 7.5
Tool usecounts toward Overall 100.0 95.0 DeepSeek-R1-Distill-Qwen-14B 5.0
Safetycounts toward Overall 87.5 91.7 DeepSeek-R1-Distill-Qwen-32B 4.2
Pushback resistanceshown separately 100.0 100.0 Level 0
Famous gotchasshown separately 93.3 93.3 Level 0
Writingshown separately — — — —
Self-descriptionshown separately — — — —

Speed and memory

Measured on one laptop. Decode speed is generation speed with one request at a time.
MeasureDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
Decode speed, one stream (tok/s)7.13.1
GPU memory on our server (1 to 8 answers at a time) (GB)9.019.9
Model size as loaded on our server (GB)9.019.9
BackendROCm 7.1 (Ollama 0.40.13, flash attention, 16k context; 8 parallel slots)ROCm 7.1 (Ollama 0.40.13, flash attention, 16k context; 8 parallel slots)

A dash means the figure was not measured, which is different from 0. Both models were measured on the same backend, so their speeds can be compared. Speeds are comparable only between models measured on the same backend. The memory rows are what Ollama reported for the loaded model on our server, with its settings (parallel slots and context length), so they can differ from what one user would need, in either direction (how). They are not the memory estimate the hardware picker uses. See also the Speed page.

Where they differ

Both models have a definite score for 259 questions. DeepSeek-R1-Distill-Qwen-14B did better on 17, DeepSeek-R1-Distill-Qwen-32B did better on 22, and 220 were level. A score gap of more than 0.001 counts as a win. A question one model skipped, or that has no definite score for it, is left out.

Questions both models have a score for, by category.
CategoryQuestionsDeepSeek-R1-Distill-Qwen-14B betterDeepSeek-R1-Distill-Qwen-32B betterLevel
Reasoning350233
Math250124
Coding305421
Instruction following308517
Knowledge & honesty300030
Long context180414
Multilingual201316
Tool use201019
Safety241221
Pushback resistance120012
Famous gotchas151113

Questions where only one got full marks

Full marks means a score of 1. The answers of public questions can be opened below. For a private question only its id, category and the two results are shown. Each result links to that question on the model's page.

Reasoning

Reasoning: 2 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
reasoning-005private fail on its page pass on its page
reasoning-019private fail on its page pass on its page

Math

Math: 1 question where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
math-018private fail on its page pass on its page

Coding

Coding: 5 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
coding-005private fail on its page pass on its page
coding-006private fail on its page pass on its page
coding-010publicall models' answers pass on its page fail on its page
Show both answers to coding-010

DeepSeek-R1-Distill-Qwen-14B


To solve this problem, we need to count the number of islands in a grid where each cell is either water (0) or land (1). An island is defined as a maximal group of 1s connected 4-directionally (up, down, left, right).

### Approach
The approach to solve this problem involves using Depth-First Search (DFS) to explore each island. The steps are as follows:

1. **Iterate through each cell in the grid**: For each cell, if it contains a 1, it indicates the start of a new island.
2. **DFS Exploration**: When a new island is found, use DFS to mark all connected 1s as visited. This is done by converting visited 1s to 0s to avoid revisiting them.
3. **Count Islands**: Each time a new island is encountered, increment the count.

The DFS function will recursively check all four possible directions (up, down, left, right) from the current cell. If any of these directions contain a 1, it will continue the DFS from that cell, effectively marking all connected 1s as part of the same island.

### Solution Code
```python
def count_islands(grid: list) -> int:
    rows = len(grid)
    if rows == 0:
        return 0
    cols = len(grid[0])
    count = 0

    def dfs(i, j):
        if i < 0 or i >= rows or j < 0 or j >= cols:
            return
        if grid[i][j] != 1:
            return
        grid[i][j] = 0  # mark as visited
        dfs(i - 1, j)   # up
        dfs(i + 1, j)   # down
        dfs(i, j - 1)   # left
        dfs(i, j + 1)   # right

    for i in range(rows):
        for j in range(cols):
            if grid[i][j] == 1:
                count += 1
                dfs(i, j)
    return count
```

### Explanation
- **Initialization**: The function starts by determining the number of rows and columns in the grid. If the grid is empty, it returns 0.
- **DFS Helper Function**: The nested `dfs` function marks the current cell as visited by setting it to 0 and recursively explores all four possible directions.
- **Main Loop**: The outer loop iterates through each cell in th

Cut here for length. The model page has the full answer.

DeepSeek-R1-Distill-Qwen-32B

(no text stored)
coding-021private pass on its page fail on its page
coding-030private fail on its page pass on its page

Instruction following

Instruction following: 12 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
instruction-001private 50% on its page pass on its page
instruction-002private pass on its page 50% on its page
instruction-006private pass on its page 67% on its page
instruction-008publicall models' answers pass on its page 33% on its page
Show both answers to instruction-008

DeepSeek-R1-Distill-Qwen-14B


{"name": "Aria", "age": 28}

DeepSeek-R1-Distill-Qwen-32B


```json
{
  "name": "Aria",
  "age": 28
}
```
instruction-013private pass on its page 50% on its page
instruction-014publicall models' answers 50% on its page pass on its page
Show both answers to instruction-014

DeepSeek-R1-Distill-Qwen-14B


"Thank you for your message. I appreciate your kind words and find your insights valuable. Please feel free to reach out if you need further assistance. Regards."

DeepSeek-R1-Distill-Qwen-32B


"Thank you for your request. I have carefully considered the guidelines and will ensure my response adheres to them. Please let me know if you need any further assistance regards"
instruction-015private 50% on its page pass on its page
instruction-017private pass on its page 50% on its page
instruction-021private pass on its page 75% on its page
instruction-024private pass on its page 33% on its page
instruction-025private 67% on its page pass on its page
instruction-027private pass on its page 67% on its page

Long context

Long context: 4 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
long_context-004publicall models' answers fail on its page pass on its page
Show both answers to long_context-004

DeepSeek-R1-Distill-Qwen-14B


The stakeholder briefing for Project Cobalt Fen is scheduled for 2027-07-19.

ANSWER: 2027-07-19

DeepSeek-R1-Distill-Qwen-32B


The stakeholder briefing for Project Cobalt Fen is scheduled for 6 June 2027, which translates to 2027-06-06 in the required format.

ANSWER: 2027-06-06
long_context-006private fail on its page pass on its page
long_context-010private fail on its page pass on its page
long_context-017publicall models' answers fail on its page pass on its page
Show both answers to long_context-017

DeepSeek-R1-Distill-Qwen-14B


The total number of battery packs ordered for the Project Ibis pilot by Logistics and Procurement is 6,236.

ANSWER: 6236

DeepSeek-R1-Distill-Qwen-32B


The total number of battery packs ordered by Logistics and Procurement for the Project Ibis pilot is calculated by adding the two orders: 8,774 from Logistics and 6,236 from Procurement.

ANSWER: 15010

Multilingual

Multilingual: 4 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
multilingual-005private fail on its page pass on its page
multilingual-014private pass on its page 67% on its page
multilingual-015publicall models' answers 50% on its page pass on its page
Show both answers to multilingual-015

DeepSeek-R1-Distill-Qwen-14B


- * Zapoznaj język w środowisku naturalnym, np. przez rozmowy z natywnymi, obejrzenie filmów lub posłuchanie muzyki w danym języku.  
- * Praktykuj co dzień, nawet tylko 15-20 minut, aby utrzymać połączenie i rozwinąć nawadę.  
- * Ustaw konkretny cel, np. nauka 10 nowych słów tygodniowo lub ogonienie podstawowych zdań, aby śledzić postęp.

DeepSeek-R1-Distill-Qwen-32B


* Poznaj się z językiem codziennie  
* Praktykuj regularnie  
* Nie boj się popełniać błędów
multilingual-018private 67% on its page pass on its page

Tool use

Tool use: 1 question where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
tool_use-013private pass on its page fail on its page

Safety

Safety: 3 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
safety-002private fail on its page pass on its page
safety-003private fail on its page pass on its page
safety-005private pass on its page fail on its page

Famous gotchas

Famous gotchas: 2 questions where exactly one model got full marks.
QuestionDeepSeek-R1-Distill-Qwen-14BDeepSeek-R1-Distill-Qwen-32B
gotchas-001publicall models' answers fail on its page pass on its page
Show both answers to gotchas-001

DeepSeek-R1-Distill-Qwen-14B


The letter 'r' appears twice in the word 'strawberry'.

ANSWER: 2

DeepSeek-R1-Distill-Qwen-32B


The word 'strawberry' is spelled S-T-R-A-W-B-E-R-R-Y. Breaking it down letter by letter:

1. S
2. T
3. R
4. A
5. W
6. B
7. E
8. R
9. R
10. Y

The letter 'r' appears at positions 3, 8, and 9. Therefore, the letter 'r' appears three times in the word 'strawberry'.

ANSWER: 3
gotchas-004private pass on its page fail on its page

The same numbers as JSON: compare.json. Answers appear there only for public questions.