Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.
GPT-5 nano is clearly ahead on the aggregate, 45 to 31. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5 nano's sharpest advantage is in mathematics, where it averages 85.2 against 40.9. The single biggest benchmark swing on the page is AIME 2025, 85.2 to 28. Ministral 3 3B (Reasoning) does hit back in multilingual, so the answer changes if that is the part of the workload you care about most.
GPT-5 nano is also the more expensive model on tokens at $0.05 input / $0.40 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Ministral 3 3B (Reasoning). That is roughly Infinityx on output cost alone. GPT-5 nano gives you the larger context window at 400K, compared with 128K for Ministral 3 3B (Reasoning).
Pick GPT-5 nano if you want the stronger benchmark profile. Ministral 3 3B (Reasoning) only becomes the better choice if multilingual is the priority or you want the cheaper token bill.
GPT-5 nano
37.7
Ministral 3 3B (Reasoning)
34
GPT-5 nano
22
Ministral 3 3B (Reasoning)
7.2
GPT-5 nano
56.7
Ministral 3 3B (Reasoning)
30.4
GPT-5 nano
58.8
Ministral 3 3B (Reasoning)
35.3
GPT-5 nano
63.7
Ministral 3 3B (Reasoning)
25.2
Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.
GPT-5 nano
48
Ministral 3 3B (Reasoning)
59.7
GPT-5 nano
85.2
Ministral 3 3B (Reasoning)
40.9
GPT-5 nano is ahead overall, 45 to 31. The biggest single separator in this matchup is AIME 2025, where the scores are 85.2 and 28.
GPT-5 nano has the edge for knowledge tasks in this comparison, averaging 63.7 versus 25.2. Inside this category, GPQA is the benchmark that creates the most daylight between them.
GPT-5 nano has the edge for coding in this comparison, averaging 22 versus 7.2. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.
GPT-5 nano has the edge for math in this comparison, averaging 85.2 versus 40.9. Inside this category, AIME 2025 is the benchmark that creates the most daylight between them.
GPT-5 nano has the edge for reasoning in this comparison, averaging 58.8 versus 35.3. Inside this category, MRCRv2 is the benchmark that creates the most daylight between them.
GPT-5 nano has the edge for agentic tasks in this comparison, averaging 37.7 versus 34. Inside this category, BrowseComp is the benchmark that creates the most daylight between them.
GPT-5 nano has the edge for multimodal and grounded tasks in this comparison, averaging 56.7 versus 30.4. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.
Ministral 3 3B (Reasoning) has the edge for multilingual tasks in this comparison, averaging 59.7 versus 48. Inside this category, MMLU-ProX 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.