NVIDIANemotron Nano 12B v2 VL
Best configuration ranks #471 of 610 on the BenchLeader Index at 42.0 ±3.1 (thinking reasoning effort). Released 28 Oct 2025.
- Blended price
- $0.300/M
- $0.200 in · $0.600 out
- Output speed
- 42 tok/s
- First answer
- 58 s
- first token 2.36 s
- Context
- 128k
- Overall index42
- Agents & tools38
- Knowledge34
- Instruction following35
- Multimodal37
- Long context46
- Composite38
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| thinkingbest | 42.0 | #471 | 42 tok/s | 58 s | $0.0003 | Agents & tools 38 · Composite 38 · Instruction following 35 · Knowledge 34 · Long context 46 · Multimodal 37 |
| default | 39.2 | #533 | 64 tok/s | 2.36 s | $0.0003 | Agents & tools 34 · Composite 36 · Instruction following 30 · Knowledge 30 · Long context 35 · Multimodal 28 · Reasoning 48 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | thinking | default | Source |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 57.2%#374 | 43.9%#430 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 5.5%#347 | 4.3%#434 | Artificial Analysis |
| Kagi LLM Benchmark | – | 51.7%#80 | Kagi LLM Benchmark |
Agents & tools
| Benchmark | thinking | default | Source |
|---|---|---|---|
| Terminal-Bench Hard | 4.5%#270 | 0.0%#345 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 21.4%#309 | 19.3%#328 | Artificial Analysis |
Knowledge
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-Omniscience | -64.2#418 | -72.2#444 | Artificial Analysis |
Instruction following
| Benchmark | thinking | default | Source |
|---|---|---|---|
| IFBench | 31.9%#330 | 25.9%#371 | Artificial Analysis |
Multimodal
| Benchmark | thinking | default | Source |
|---|---|---|---|
| MMMU-Pro | 52.9%#200 | 44.5%#221 | Artificial Analysis |
Long context
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA-LCR | 41.0%#286 | 20.7%#359 | Artificial Analysis |
Composite
| Benchmark | thinking | default | Source |
|---|---|---|---|
| AA Intelligence Index | 7.5#393 | 5.8#469 | Artificial Analysis |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0003 | – | 1.1 min |
| Summarise a 30-page report | 12,000 / 600 | $0.0028 | – | 1.2 min |
| Code edit | 6,000 / 1,500 | $0.0021 | – | 1.6 min |
| Agentic coding session | 60,000 / 4,000 | $0.014 | – | 2.5 min |
| Structured extraction | 2,000 / 200 | $0.0005 | – | 1.1 min |
See also
Data as of 9 Sept 2026. Compare these configurations.