Nemotron 3.5 Super VL 120B A12B
Nemotron 3.5 Super VL 120B A12BNVIDIAOpen weightsReasoningProvisionalReleased Oct 15, 2026
Overall score
49.4
Rank 102 of 222 · 35.0–63.7
Context window
1M
1,000,000 tokens
Price per 1M tokens
— in · — out
No price listed
Blended price
—
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Reasoning72.1
- 2Instruction following#7466.6
- 3Agentic & tool use64.0
- 4Multimodal62.0
- 5Knowledge47.1
- 6Coding40.2
Benchmark results (17)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Multimodal
| CharXiv | 80.8 | #25 of 42 |
| MathVision | 78.2 | #19 of 20 |
| V* | 85.6 | #10 of 12 |
| BabyVision | 29.6 | #8 of 8 |
| MMLongBench-Doc | 62.8 | #1 of 5 |
Agentic & tool use
| Terminal-Bench 2.1 | 66.7 | #45 of 68 |
| BrowseComp | 78.0 | #24 of 50 |
| τ³-bench results | 18.9 | #19 of 23 |
| τ²-bench Airline | 76.5 | #2 of 4 |
Coding
| SWE-bench Verified | 71.1 | #58 of 98 |
| SWE-bench Pro | 51.8 | #58 of 69 |
| LiveCodeBench v6 | 88.2 | #11 of 46 |
Knowledge
| HLE | 29.2 | #45 of 69 |
| HLE w/o tools | 31.8 | #28 of 43 |
Multilingual
| SWE Multilingual | 67.7 | #36 of 46 |
Instruction following
| IFBench | 70.9 | #29 of 42 |
Math
| Apex Shortlist | 88.4 | #3 of 4 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| List price | — | — | — |
NVIDIA publishes downloadable OpenMDW 1.1 weights. No exact first-party per-token hosted API tariff is established for this checkpoint; hardware costs are separate.
Speed
Measured by us through OpenRouter (median of trials).
Not available on OpenRouter, so we cannot measure it yet.
Related models
Same family
Similar score
- Kimi K2.7 Code49.4 · $1.71
- Gemini 3.5 Flash-Lite49.6 · $0.85
- Gemini 2.5 Pro49.1 · $3.44
- GLM-5V-Turbo49.6 · $1.90
- Qwen3.5 Plus48.9 · $0.90