NVIDIA Nemotron Nano 12B v2 VL
Oct 2025
NVIDIA Nemotron Nano 12B v2 VLCompare inon
Benchmarks of NVIDIA Nemotron Nano 12B v2 VL14
BenchmarkByScoreRank
Overall1
Knowledge4
Science2
Coding1
Multimodal1
Documents1
Legacy · saturated or no longer reported4
No benchmark matches
Rankingin non-thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperClaude Haiku 5.5max×7.5-33%$0.21Remove sampling parameters+5
BetterGPT-6 Lunalow×3.7-33%$0.005Function calling only with none · Chat Completions+3
OpenQwen3.8 27Bxhigh×5.8+219%$1.01Memory (4-bit): 6.6 GB → 17 GB+5
CheaperMiniMax M3.1 Flashthinking×7.3-13%$0.22Complete model response must be appended
CheaperGLM 5.3 FlashmaxOpen×7.2-21%$0.28Memory (4-bit): 6.6 GB → 193 GB+5ModelScoreCostPer taskWhat changes
Claude Haiku 5.5max×7.5-33%$0.21Remove sampling parameters+5MiniMax M3.1 Flashthinking×7.3-13%$0.22Complete model response must be appendedGLM 5.3 FlashmaxOpen×7.2-21%$0.28Memory (4-bit): 6.6 GB → 193 GB+5Qwen3.8 FlashthinkingOpen×6.8-23%$0.37Memory (4-bit): 6.6 GB → 108 GB+6GPT-6 Lunamax×6.6-33%$0.068Function calling only with none · Chat Completions+3ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×3.7-33%$0.005Function calling only with none · Chat Completions+3
Granite 4.2 3BthinkingOpen+56%-83%$0.006No image input
Gemma 4 E4BthinkingOpen+53%-87%$0.008GPT-5.6 Lunalow×3.6+50%$0.01GPT OSS 20BhighOpen+54%-88%$0.011Memory (4-bit): 6.6 GB → 13 GB+1ModelScoreCostPer taskWhat changes
Qwen3.8 27BxhighOpen×5.8+219%$1.01Memory (4-bit): 6.6 GB → 17 GB+5Qwen 3.5 27BthinkingOpen×3.9+175%$0.30 / $2.40Memory (4-bit): 6.6 GB → 17 GBQwen3.6 27BthinkingOpen×3.7+350%$0.62Memory (4-bit): 6.6 GB → 17 GBIK2 Horizon 7BthinkingOpen×3.5+75%$0.30 / $1.20No image input
Muse Glimmer 30BhighOpen×3+67%$0.058Memory (4-bit): 6.6 GB → 18 GBModelScoreCostPer taskWhat changes
Nemotron 3.5 Super VLOpen×4.6—$0.19Memory (4-bit): 6.6 GB → 75 GBNemotron 3 UltrathinkingOpen×3.9—$0.58Memory (4-bit): 6.6 GB → 336 GB+1Nemotron 3.5 Bolt 74B-A9BOpen×3.7——Memory (4-bit): 6.6 GB → 45 GBNemotron 3.5 LightningthinkingOpen×2.2—$0.079Memory (4-bit): 6.6 GB → 20 GB+1Nemotron 3 Super 120B A12BOpen×2.2—$1.64Memory (4-bit): 6.6 GB → 74 GB+1Nemotron Cascade 2 30B A3BOpen×2——Memory (4-bit): 6.6 GB → 19 GB+1Nemotron 3 Nano Omni 30B A3B ReasoningOpen+76%—$0 / $0Memory (4-bit): 6.6 GB → 20 GBLlama Nemotron Super 49B v1.5non-thinkingOpen+27%—$0.40 / $0.40Memory (4-bit): 6.6 GB → 30 GB+1All NVIDIA models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · 8-bit✓ 24 GB GPU · 8-bit✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
Licencenvidia-open-model-licensecommercial use with conditions
Benchmarks
all categoriesChat & Knowledge
Humanity's Last Exam4.3#300
AA Omniscience Index-72.2#299
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈5.8#300
AA Omniscience · Accuracy11.9#268
AA-LCR (long context) 1.120.7#240
GPQA Diamond43.9#270
AIME 2025 · AA archived26.7#108
Coding
LiveCodeBench · AA archived34.5#111
Agents & Tool Use
τ²-bench · AA19.3#199
Writing & Language
IFBench (instructions)25.9#228
Vision & Documents
MMMU-Pro (vision)44.5#118
FAQ
How much does NVIDIA Nemotron Nano 12B v2 VL cost?›
$0.20 per 1M input tokens and $0.60 per 1M output tokens (cheapest host: DeepInfra).
Is NVIDIA Nemotron Nano 12B v2 VL free to use?›
There is no free API; the cheapest paid price is $0.20 per 1M input tokens. The weights are open, so you can run it yourself.
How good is NVIDIA Nemotron Nano 12B v2 VL, and what is it best at?›
#297 of 334 in Chat & Knowledge, #209 of 239 in Coding, #197 of 242 in Agents & Tool Use, #225 of 245 in Writing & Language, #115 of 124 in Vision & Documents.
Is NVIDIA Nemotron Nano 12B v2 VL better than Solar Pro 2?›
On AA Intelligence Index 4.3.2, NVIDIA Nemotron Nano 12B v2 VL scores 7.5 vs 7.5 — about the same.
How can I use NVIDIA Nemotron Nano 12B v2 VL?›
Through the API of DeepInfra.
What is the context window of NVIDIA Nemotron Nano 12B v2 VL?›
128K tokens.
When was NVIDIA Nemotron Nano 12B v2 VL released? Is there a newer version?›
Released 2025-10-28. It is the current version in its line.
Is NVIDIA Nemotron Nano 12B v2 VL open source?›
Yes, the weights are open (nvidia-open-model-license licence).
Can I run NVIDIA Nemotron Nano 12B v2 VL locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can NVIDIA Nemotron Nano 12B v2 VL read images?›
It accepts text and images as input.
What is NVIDIA Nemotron Nano 12B v2 VL?›
NVIDIA Nemotron Nano 12B v2 VL is an open-weights language model from NVIDIA, released Oct 2025.
What are the best alternatives to NVIDIA Nemotron Nano 12B v2 VL?›
Claude Haiku 5.5 (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).