Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
DeepSeek V3.1 (Reasoning) is clearly ahead on the aggregate, 41 to 36. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
DeepSeek V3.1 (Reasoning)'s sharpest advantage is in multimodal & grounded, where it averages 41.5 against 33.4. The single biggest benchmark swing on the page is LongBench v2, 57 to 44. Ministral 3 8B (Reasoning) does hit back in mathematics, so the answer changes if that is the part of the workload you care about most.
Pick DeepSeek V3.1 (Reasoning) if you want the stronger benchmark profile. Ministral 3 8B (Reasoning) only becomes the better choice if mathematics is the priority.
DeepSeek V3.1 (Reasoning)
44.2
Ministral 3 8B (Reasoning)
38.5
DeepSeek V3.1 (Reasoning)
19.9
Ministral 3 8B (Reasoning)
15.2
DeepSeek V3.1 (Reasoning)
41.5
Ministral 3 8B (Reasoning)
33.4
DeepSeek V3.1 (Reasoning)
45.8
Ministral 3 8B (Reasoning)
42.1
DeepSeek V3.1 (Reasoning)
32.8
Ministral 3 8B (Reasoning)
30
DeepSeek V3.1 (Reasoning)
70
Ministral 3 8B (Reasoning)
70
DeepSeek V3.1 (Reasoning)
62.1
Ministral 3 8B (Reasoning)
61.7
DeepSeek V3.1 (Reasoning)
46.1
Ministral 3 8B (Reasoning)
47.8
DeepSeek V3.1 (Reasoning) is ahead overall, 41 to 36. The biggest single separator in this matchup is LongBench v2, where the scores are 57 and 44.
DeepSeek V3.1 (Reasoning) has the edge for knowledge tasks in this comparison, averaging 32.8 versus 30. Inside this category, HLE is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for coding in this comparison, averaging 19.9 versus 15.2. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.
Ministral 3 8B (Reasoning) has the edge for math in this comparison, averaging 47.8 versus 46.1. Inside this category, MATH-500 is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for reasoning in this comparison, averaging 45.8 versus 42.1. Inside this category, LongBench v2 is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for agentic tasks in this comparison, averaging 44.2 versus 38.5. Inside this category, OSWorld-Verified is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) has the edge for multimodal and grounded tasks in this comparison, averaging 41.5 versus 33.4. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.
DeepSeek V3.1 (Reasoning) and Ministral 3 8B (Reasoning) are effectively tied for instruction following here, both landing at 70 on average.
DeepSeek V3.1 (Reasoning) has the edge for multilingual tasks in this comparison, averaging 62.1 versus 61.7. Inside this category, MGSM is the benchmark that creates the most daylight between them.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.