Model vs model
Mistral Small 4 vs NVIDIA Nemotron 3.5 Super VL
Both are open-weights models priced almost the same, at $0.28 and $0.26 per 1M tokens blended, and both score 0 on Terminal-Bench 4.0. NVIDIA Nemotron 3.5 Super VL is the better pick for hard reasoning and long documents, with an Intelligence Index of 26.5 against 11.3 and a 1.05M-token context window. Mistral Small 4 suits jobs that need PDF input, since it accepts PDFs and NVIDIA Nemotron 3.5 Super VL does not.
At a glance
Longer bar means the bigger number.
Blended price · per 1M tokens, lower is better
Intelligence Index · Artificial Analysis, higher is better
Context window · tokens it can read at once
Where they sit among all 44 models
Mistral Small 4 ranks 39th of the 43 models with an Intelligence Index, NVIDIA Nemotron 3.5 Super VL 25th. Up and to the left is smarter for less money. Hover or tap a dot for its numbers, then click it to compare that model with Mistral Small 4.
Which one to pick
Pick Mistral Small 4 if
- you want the lower bill ($0.26 vs $0.28 per 1M tokens blended)
- you need to send PDFs
Pick NVIDIA Nemotron 3.5 Super VL if
- you want the higher overall score (26.5 vs 11.3)
- you care about Humanity's Last Exam (hard reasoning): 30.7 vs 9.9
- you care about AA-LCR (long documents): 69 vs 49.7
- you need a longer context window (1.05M vs 262K tokens)
How smart is each one?
Of the 11 scores both models have, NVIDIA Nemotron 3.5 Super VL is ahead on 10, and one is within a point.
Every intelligence score we have
Out of 100, higher is better. Mistral Small 4 NVIDIA Nemotron 3.5 Super VL
Overall
Intelligence Index
11.326.5
By area
Coding
26.6NVIDIA Nemotron 3.5 Super VL: not scored yet
Agentic work
0.8NVIDIA Nemotron 3.5 Super VL: not scored yet
By field of work
Finance & accounting
10.527.2Legal
12.828.1Healthcare & medicine
10.825.5Economics
18.138.5Engineering
16.430.1Strategy & operations
6.722.6
Benchmarks
Hard reasoning · Humanity's Last Exam
9.930.7Science · GPQA Diamond
76.9NVIDIA Nemotron 3.5 Super VL: not scored yet
Scientific coding · SciCode
38.847.8Terminal tasks · Terminal-Bench 4.0
0 eachTool use · τ-Bench Banking
4.9NVIDIA Nemotron 3.5 Super VL: not scored yet
Instruction following · IFBench
48.2NVIDIA Nemotron 3.5 Super VL: not scored yet
Long documents · AA-LCR
49.769Uncensored · UGI
37.4NVIDIA Nemotron 3.5 Super VL: not scored yet
The specs
| Spec | Mistral Small 4 | NVIDIA Nemotron 3.5 Super VL |
|---|---|---|
| Input price | $0.15 per 1M | $0.20 per 1M |
| Output price | $0.60 per 1M | $0.50 per 1M |
| Cost per test task | $0.0150 | Not tested |
| Context window | 262K | 1.05M |
| Max output | 256K | 1.05M |
| Accepts | Text, Image, PDF | Text, Image |
| Open weights | Yes | Yes |
| Lab | Mistral AI | NVIDIA |
| Released | Mar 2026 | Oct 2026 |
| Knowledge cutoff | Jun 2025 | Not published |
Blended price weighs 3 parts input to 1 part output. List prices from models.dev. Cost per test task from Artificial Analysis.
Questions
Is Mistral Small 4 or NVIDIA Nemotron 3.5 Super VL cheaper?
Mistral Small 4 is cheaper. It charges $0.15 per 1M input tokens and $0.60 per 1M output tokens, against $0.20 and $0.50 for NVIDIA Nemotron 3.5 Super VL. Blended 3 parts input to 1 part output, that is $0.26 against $0.28 per 1M tokens.
Which is smarter, Mistral Small 4 or NVIDIA Nemotron 3.5 Super VL?
NVIDIA Nemotron 3.5 Super VL scores higher on the Artificial Analysis Intelligence Index, 26.5 against 11.3 for Mistral Small 4. The biggest differences are on Humanity's Last Exam (9.9 vs 30.7) and AA-LCR (49.7 vs 69).
Which has the longer context window?
NVIDIA Nemotron 3.5 Super VL accepts 1.05M tokens, against 262K for Mistral Small 4.
Can I compare Mistral Small 4 and NVIDIA Nemotron 3.5 Super VL on my own prompt?
Yes. ModelCompare sends one prompt to both models at once and shows each answer side by side with what it cost. You can try it free without an account.
Related comparisons
Where the numbers come from
Intelligence Index and benchmark scores are from Artificial Analysis, at each model’s highest effort, except SimpleQA, which is from Epoch AI, and the Uncensored score, which is from the UGI Leaderboard. The hallucination test is ours: questions with one exact answer, fake evidence pushed against right answers and requests that tempt a model to name packages that don’t exist. Prices are list prices from models.dev. Example answers come from our own test runs. Updated October 25, 2026.