Nemotron 3 Super 120B A12B Official Specifications, Pricing & Benchmarks
Nemotron 3 Super 120B A12B (nemotron-3-super-120b-a12b) by NVIDIA, released 2026-03-11, featuring 262,144 context window, 16,384 max output tokens, and Text input.
📋 Nemotron 3 Super 120B A12B Official API & System Card Specifications
Verified against official NVIDIA developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- nemotron-3-super-120b-a12b
- Release Date
- 2026-03-11
- Weight License
- Open-Weight (API Spec)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- ✅ 原生支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 262,000 tokens (262K)
- Max Output Tokens
- 16,384 tokens
- Input Modalities
- Text
- Output Modalities
- Text
- Direct China Access
- 🌐 Global / OpenRouter
- Deployment Mode
- Open-Weight + Cloud API
3. Official Pricing & Cache
USD / 1M- Input / 1M Tokens
- $ 0.075
- Cached Input / 1M
- $ 0.015 -80%
- Output / 1M Tokens
- $ 0.4
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
📊 Nemotron 3 Super 120B A12B Full Benchmark Scorecard
23 public benchmarks · ranks among 154 models · synced 2026-11-20
💻 Coding 6 items
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code.
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
A multilingual benchmark for issue resolving in software engineering that covers Java, TypeScript, JavaScript, Go, Rust, C, and C++.
SciCode is a research coding benchmark curated by scientists that challenges language models to code solutions for scientific problems.
Terminal-Bench 2.0 is an updated benchmark for testing AI agents' tool use ability to operate a computer via terminal.
Terminal-Bench is a benchmark for testing AI agents in real terminal environments. It evaluates how well agents can handle real-world, end-to-end tasks autonomously, including compiling code, training models, setting up servers, system administration, security tasks, data…
🤖 Agents & Tools 4 items
τ²-Bench telecom domain evaluates conversational agents in a dual-control environment modeled as a Dec-POMDP, where both agent and user use tools in shared telecommunications troubleshooting scenarios that test coordination and communication capabilities.
τ²-bench retail domain evaluates conversational AI agents in customer service scenarios within a dual-control environment where both agent and user can interact with tools.
TAU2 airline domain benchmark for evaluating conversational agents in dual-control environments where both AI agents and users interact with tools in airline customer service scenarios.
BrowseComp is a benchmark comprising 1,266 questions that challenge AI agents to persistently navigate the internet in search of hard-to-find, entangled information.
📐 Math 5 items
Harvard-MIT Mathematics Tournament 2025 - A prestigious student-organized mathematics competition for high school students featuring two tournaments (November 2025 at MIT and February 2026 at Harvard) with individual tests, team rounds, and guts rounds.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks.
Extended version of MMLU-Pro providing additional challenging multiple-choice questions for evaluating language models across diverse academic and professional domains. Built on the foundation of the Massive Multitask Language Understanding benchmark framework.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions.
📜 Long Context 2 items
RULER v1 is a synthetic long-context benchmark for measuring how model quality degrades as input length increases. This packaging follows the public standalone NVIDIA RULER implementation with 13 official tasks spanning retrieval, multi-hop tracing, aggregation, and QA.
Agent Arena Long Context Reasoning benchmark.
🧠 Reasoning & Science 4 items
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
Arena-Hard-Auto v2 is a challenging benchmark consisting of 500 carefully curated prompts sourced from Chatbot Arena and WildChat-1M, designed to evaluate large language models on real-world user queries.
MultiChallenge is a realistic multi-turn conversation evaluation benchmark that challenges frontier LLMs across four key categories: instruction retention (maintaining instructions throughout conversations), inference memory (recalling and connecting details from previous…
BIRD (BIg Bench for LaRge-scale Database Grounded Text-to-SQLs) is a comprehensive text-to-SQL benchmark containing 12,751 question-SQL pairs across 95 databases (33.4 GB total) spanning 37+ professional domains.
Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).
Closest Self-Hostable Open-Weight Alternatives to Nemotron 3 Super 120B A12B
Open-weight models with full HuggingFace weights for private GPU / Mac deployment
Frequently Asked Questions
Is Nemotron 3 Super 120B A12B open-source or proprietary?
Nemotron 3 Super 120B A12B (nemotron-3-super-120b-a12b) is a proprietary commercial model by NVIDIA. Weights and internal layer dimensions are undisclosed; it is accessed via cloud API.
What are the context limit and API pricing of Nemotron 3 Super 120B A12B?
Nemotron 3 Super 120B A12B supports 262K input context and 16,384 max output tokens, priced at $0.075/1M input ($0.015/1M cached) and $0.4/1M output.