The AI arena is free today

Open Superagent

Inkling-Small vs Nemotron 3 Ultra (550B A55B)

Inkling-Small and Nemotron 3 Ultra (550B A55B) are closely matched at 39.3 and 37.0 on the LLM Stats Score.

Thinking Machines Lab · NVIDIA · Updated for 2026

Which is better?

Inkling-Small and Nemotron 3 Ultra (550B A55B) are closely matched on the overall LLM Stats Score at 39.3 and 37.0.

In the 8 individual benchmarks reported for both models, Inkling-Small wins 7; this is a narrower head-to-head signal than the composite indexes.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Inkling-Small

  • your work emphasizes coding and agents — it leads those capability indexes
  • you value its reported benchmark strengths — it wins 7 of 8 exact shared results
  • you want the most recent training data — it shipped Jul 2026

Choose Nemotron 3 Ultra (550B A55B)

  • you are already invested in the NVIDIA ecosystem

At a glance

The differences that matter most.

Inkling-Small
Thinking Machines Lab
Nemotron 3 Ultra (550B A55B)
NVIDIA
Core performance indexes
LLM Stats Score
39.3
#53
37.0
#63
Reasoning
39.0
#52
36.3
#67
Coding
31.1
#45
22.1
#72
Agents
24.8
#45
13.4
#86
Cost, coverage & limits
Benchmark wins
7 of 8
1 of 8
Input price
$0.30 / M
— / M
Output price
$1.20 / M
— / M
Context window
256,000
—

Capability indexes

Additional strengths measured across groups of related public benchmarks

2 shared
Index
Inkling-Small
Nemotron 3 Ultra (550B A55B)
Math
34.2#46
36.4#37
Tool use
23.3#38
7.4#132
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

24 reported for Inkling-Small · 26 for Nemotron 3 Ultra (550B A55B)

8 shared

Inkling-Small outperforms in 7 benchmarks (BrowseComp, CritPT, GPQA, IFBench, SciCode, SWE-Bench Verified, Terminal-Bench 2.1), while Nemotron 3 Ultra (550B A55B) is better at 1 benchmark (Humanity's Last Exam).

Inkling-Small significantly outperforms across most benchmarks.

Fri Aug 28 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

274.0B diff

Nemotron 3 Ultra (550B A55B) has 274.0B more parameters than Inkling-Small, making it 99.3% larger.

Thinking Machines Lab
Inkling-Small
276.0Bparameters
NVIDIA
Nemotron 3 Ultra (550B A55B)
550.0Bparameters
276.0B
Inkling-Small
550.0B
Nemotron 3 Ultra (550B A55B)

Context Window

Maximum input and output token capacity

Only Inkling-Small specifies input context (256,000 tokens). Only Inkling-Small specifies output context (256,000 tokens).

Thinking Machines Lab
Inkling-Small
Input256,000 tokens
Output256,000 tokens
NVIDIA
Nemotron 3 Ultra (550B A55B)
Input- tokens
Output- tokens
Fri Aug 28 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Inkling-Small supports multimodal inputs, whereas Nemotron 3 Ultra (550B A55B) does not.

Inkling-Small can handle both text and other forms of data like images, making it suitable for multimodal applications.

Inkling-Small

Text
Images
Audio
Video

Nemotron 3 Ultra (550B A55B)

Text
Images
Audio
Video

License

Usage and distribution terms

Inkling-Small is licensed under Apache 2.0, while Nemotron 3 Ultra (550B A55B) uses OpenMDW License v1.1.

License differences may affect how you can use these models in commercial or open-source projects.

Inkling-Small

Apache 2.0

Open weights

Nemotron 3 Ultra (550B A55B)

OpenMDW License v1.1

Open weights

Release Timeline

When each model was launched

Inkling-Small was released on 2026-07-30, while Nemotron 3 Ultra (550B A55B) was released on 2026-06-04.

Inkling-Small is 2 months newer than Nemotron 3 Ultra (550B A55B).

Inkling-Small

Jul 30, 2026

4 weeks ago

1mo newer
Nemotron 3 Ultra (550B A55B)

Jun 4, 2026

2 months ago

Knowledge Cutoff

When training data ends

Nemotron 3 Ultra (550B A55B) has a documented knowledge cutoff of 2025-09-30, while Inkling-Small's cutoff date is not specified.

We can confirm Nemotron 3 Ultra (550B A55B)'s training data extends to 2025-09-30, but cannot make a direct comparison without Inkling-Small's cutoff date.

Inkling-Small

—

Nemotron 3 Ultra (550B A55B)

Sep 2025

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Judge for yourself.

Run your own prompts against Inkling-Small and Nemotron 3 Ultra (550B A55B) side-by-side, then vote on the output you prefer.

Inkling-Small
✓ Preferred
Nemotron 3 Ultra (550B A55B)
Open in Playground

FAQ

Common questions about Inkling-Small vs Nemotron 3 Ultra (550B A55B).

Which is better, Inkling-Small or Nemotron 3 Ultra (550B A55B)?

Inkling-Small and Nemotron 3 Ultra (550B A55B) are closely matched on the LLM Stats Score at 39.3 and 37.0. Inkling-Small is made by Thinking Machines Lab and Nemotron 3 Ultra (550B A55B) is made by NVIDIA. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Inkling-Small compare to Nemotron 3 Ultra (550B A55B) in benchmarks?

Inkling-Small scores AIME 2026: 95.5%, VoiceBench Avg: 90.1%, GPQA: 89.5%, Global-MMLU-Lite: 86.7%, ARC-AGI: 84.0%. Nemotron 3 Ultra (550B A55B) scores RULER: 94.7%, IMO-AnswerBench: 92.3%, PinchBench: 90.0%, LiveCodeBench v6: 89.0%, GPQA: 87.0%.

What are the context window sizes for Inkling-Small and Nemotron 3 Ultra (550B A55B)?

Inkling-Small supports 256K tokens and Nemotron 3 Ultra (550B A55B) supports an unknown number of tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Inkling-Small and Nemotron 3 Ultra (550B A55B)?

Key differences include LLM Stats Score (39.3 vs 37.0), multimodal support (yes vs no), licensing (Apache 2.0 vs OpenMDW License v1.1). See the full comparison above for benchmark-by-benchmark results.

Who makes Inkling-Small and Nemotron 3 Ultra (550B A55B)?

Inkling-Small is developed by Thinking Machines Lab and Nemotron 3 Ultra (550B A55B) is developed by NVIDIA.

Related comparisons