LANGUAGE MODEL Nvidia Last updated:

Llama 3.3 Nemotron Super 49B v1 (Reasoning)

Llama 3.3 Nemotron Super 49B v1 (Reasoning) is an API model from Nvidia. It’s positioned for general text tasks—work that benefits from iteration, not just one-shot answers.

API Docs ↗
Intelligence
Below avg
Context
128K
Up to 128,000 tokens
How are Intelligence, Speed & Cost bucketed?
Intelligence and Speed buckets are percentile ranks on Artificial Analysis. Cost buckets are fixed dollar thresholds keyed off output-token price ($/M out).
Intelligence
  • Top 1%≤ 1%
  • Top 5%≤ 5%
  • Top 10%≤ 10%
  • Good≤ 25%
  • Medium≤ 50%
  • Below avg> 50%
Speed
  • Top 1%≥ 345 tok/s
  • Top 5%≥ 237 tok/s
  • Top 10%≥ 196 tok/s
  • Good≥ 146 tok/s
  • Medium≥ 90 tok/s
  • Slow< 90 tok/s
Cost
  • Freeopen weights · self-host
  • Low< $1 / M out
  • Moderate$1–5 / M out
  • High≥ $5 / M out
Artificial Analysis ↗ Hugging Face ↗

Core Capabilities

Long Documents
Handles entire codebases, books, and multi-doc RAG.

Context Window

128k tokens
≈ 98 pages
4k Chat 聊天
32k Long docs 长文档
128k This model 本模型
400k Multi-doc 多文档
1M Codebase 整个代码库
10M

Availability

API
Available
Product / App
Not available
Open Source
Not released
Enterprise
Contact sales

Pricing Model

Pay per token
Input and output billed separately.
Pay-per-token

Capability / Performance

Where this model sits relative to the middle 60% of models in the tree. All scores are 0–10 (higher is better).

Lower 20% Upper 80% This model
Reasoning
AA Intelligence Index · scaled to 10
1.7
5.6
2.6
Coding
SciCode · scaled to 10
1.8
4.3
2.8
Agentic tasks
Terminal-Bench Hard · scaled to 10
0.2
3.6
0.0
Context / memory
Context window size · log-scaled
6.0
9.0
6.0
Lower 20% 20th percentile — 20% of models score below this This model Where the current model lands Upper 80% 80th percentile — only 20% of models score above this Percentile boundaries are computed across every model in the tree that reports the underlying benchmark for each capability.

What it feels like

  • Same architecture as 3.1 70B but post-training advances delivered notable reasoning + math gains
  • 92.1% on IFEval — beats Llama 3.1 405B (88.6) and GPT-4o (84.6) on instruction following
  • MATH 77.0% (vs 67.8% for 3.1 70B); HumanEval 88.4%; MMLU Chat 86.0%
Reviews: DataCamp — Llama 3.3 70B review ↗ · Artificial Analysis — Llama 3.3 70B ↗ · Analytics Vidhya — Llama 3.3 70B is here, 25x cheaper ↗

Best use cases

  • Open-weights production where 405B is too expensive but 8B is too weak
  • Self-hosted RAG and code-completion stacks needing strong instruction-following
  • Multilingual reasoning (MGSM 91.1%) at frontier-tier quality

Tools to try

NIM Microservices Build catalog

Not ideal for

  • Frontier reasoning by mid-2025 — Llama 4 family and competitors moved past it
  • Multimodal tasks (text-only; vision lives in Llama 3.2 Vision / Llama 4)