Skip to main content
Meo

NVIDIA: Nemotron Nano 12B 2 VL (free)

NVIDIA: Nemotron Nano 12B 2 VL (free) is nvidia's model. On the un-leaked meo benchmark it ranks #76 of 82 with 42.3% accuracy and an Effective Value (𝕍) of 0.0001 β€” earning a correct answer for about $0.0000 in 15s each.

as of 2026-10-18

meo rank
#76
of 82
Accuracy
42.3%
overall
𝕍 index (top=100)
0.002
raw 𝕍 0.0001 Β· N=10
$ / correct
$0.0000
15s each

Per-domain accuracy

106 correct of 250 questions asked across 11 domains Β· answer rate 100% Β· macro accuracy 28.2%. Headline accuracy is micro-averaged per meo’s methodology; per-domain question counts (correct/asked) are shown so you can judge significance.

State tracking
100%1/1
Base-rate bias
68%28/41
Framework bias
67%2/3
Perceptual illusions
56%57/102
Watson-Glaser
50%2/4
CSP (unsatisfiable)
25%4/16
Regex automaton
25%4/16
Logic / math / CS
24%8/34
Tape machine
0%0/15
Theory of mind
0%0/3
Long arithmetic
0%0/15

Efficiency

Cost per correct
$0.0000
Seconds per correct
15
Tokens per correct
465
Correct answers
106 / 250
Total benchmark cost
$0.00
Avg tokens / item
197

Model details

Released
2025-10-28
Weights
Open
Context
128K
Max output
128K
Knowledge cutoff
β€”
Price in
$0.00/M
Price out
$0.00/M
Modality
text+image+video->text
View on OpenRouter β†—Hugging Face β†—

Sources: openrouter Β· meo

FAQ

How accurate is NVIDIA: Nemotron Nano 12B 2 VL (free)?

NVIDIA: Nemotron Nano 12B 2 VL (free) scores 42.3% overall on the meo benchmark (rank #76 of 82), an un-leaked private holdout graded objective-first.

What is NVIDIA: Nemotron Nano 12B 2 VL (free)'s Effective Value?

Its Effective Value (𝕍) is 0.0001 at N=10, Ο‰=1, Ξ΄=1.5 β€” a fused metric of accuracy, speed, cost, and error-cascade. 𝕍 rankings shift with task depth.

How much does NVIDIA: Nemotron Nano 12B 2 VL (free) cost per correct answer?

About $0.0000 per correct answer (465 tokens, 15 seconds each) at benchmark-time pricing via OpenRouter.

Is NVIDIA: Nemotron Nano 12B 2 VL (free) open weights?

Yes β€” NVIDIA: Nemotron Nano 12B 2 VL (free) is released with open weights.

Benchmark updates affecting this model

  • 2026-06-09majorlogic_math_cs ground truth was broken β€” culled and rebuilt with solver-verified items

    A 5-lab frontier jury revealed that 29 of 37 LLM-authored logic_math_cs items had broken ground truth: on most, five independent frontier models unanimously concluded the clue sets were contradictory while the stored answer claimed a number, and on eight, all five agreed on the SAME different number (author arithmetic errors). Models were being marked wrong for being right. The 29 were retired (full audit published in the repo) and the domain rebuilt with generator-as-oracle items: an embedded constraint solver PROVES each puzzle has exactly one solution (or provably none), verified by an independent exhaustive enumerator. A 3-round difficulty study found frontier models saturate consistent human-scale logic puzzles, so items use anonymous global constraints and carry a measured difficulty gradient.

  • 2026-06-08majorCorrected base_bias grading (LLM-equivalence judge fix)

    The LLM-equivalence grader uses a reasoning model whose verdict token budget was too small, so its YES/NO answer was emitted as empty and silently counted as a failure. This understated the base_bias domain to ~7% for every model when the true accuracy is ~85-90%. We raised the judge budget, excluded its reasoning tokens from the verdict, switched to verbosity-robust conclusion matching, and re-graded the whole board.

Full benchmark changelog β†’

Embed this ranking

Show NVIDIA: Nemotron Nano 12B 2 VL (free)’s meo rank on your site β€” the badge links back here.

Related: all NVIDIA models Β· Multi-step state tracking domain Β· best Effective Value Β· biggest movers Β· methodology Β· full leaderboard.