Agentic
Like-for-like- Inkling
- 69.4
- Nemotron 3 Ultra
- 51.3
- Weighted basis
- 2 vs 2 rows
- Reading
- Inkling leads
Inkling vs Nemotron 3 Ultra
Updated August 29, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Inkling has the higher public score, 67.02 versus 46.32, and the 90% score intervals do not overlap.
9 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Agentic work
Tool use, computer use, and multi-step task completion
Inkling
Inkling leads on the same 2 weighted benchmark rows.
Confidence: limited
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
6 Inkling only
9 Shared
10 Nemotron 3 Ultra only
- Shared results
- 9
- Inkling only
- 6
- Nemotron 3 Ultra only
- 10
- Like-for-like categories
- 2 / 8
2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.
Instruction following
Like-for-like- Inkling
- 79.8
- Nemotron 3 Ultra
- 81.7
- Weighted basis
- 1 vs 1 rows
- Reading
- Nemotron 3 Ultra leads
Coding
Directional only- Inkling
- 68.6
- Nemotron 3 Ultra
- 58.3
- Weighted basis
- 2 vs 2 rows
- Reading
- Directional only
Knowledge
Directional only- Inkling
- 51.6
- Nemotron 3 Ultra
- 53.8
- Weighted basis
- 2 vs 3 rows
- Reading
- Directional only
Reasoning
Not comparable- Inkling
- Not measured
- Nemotron 3 Ultra
- 61.9
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Math
Not comparable- Inkling
- 97.1
- Nemotron 3 Ultra
- Not measured
- Weighted basis
- 1 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- Inkling
- Not measured
- Nemotron 3 Ultra
- 83.0
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Multimodal
Not comparable- Inkling
- 76.5
- Nemotron 3 Ultra
- Not measured
- Weighted basis
- 2 vs 0 rows
- Reading
- Not comparable
| Category | Inkling | Nemotron 3 Ultra | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 69.4 | 51.3 | Like-for-like2 vs 2 rows | Inkling leads |
| Instruction following | 79.8 | 81.7 | Like-for-like1 vs 1 rows | Nemotron 3 Ultra leads |
| Coding | 68.6 | 58.3 | Directional only2 vs 2 rows | Directional only |
| Knowledge | 51.6 | 53.8 | Directional only2 vs 3 rows | Directional only |
| Reasoning | Not measured | 61.9 | Not comparable0 vs 1 rows | Not comparable |
| Math | 97.1 | Not measured | Not comparable1 vs 0 rows | Not comparable |
| Multilingual | Not measured | 83.0 | Not comparable0 vs 1 rows | Not comparable |
| Multimodal | 76.5 | Not measured | Not comparable2 vs 0 rows | Not comparable |
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
BrowseComp
Agentic
HLE
Knowledge
Terminal-Bench 2.0
Agentic
SWE-bench Verified
Coding
IFBench
Instruction following
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- Inkling
- $0.00421
- Fits in one request
- Nemotron 3 Ultra
- Self-hosted; infrastructure cost varies
- Fits in one request
Nemotron 3 Ultra has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Inkling
- $0.10754
- Fits in one request
- Nemotron 3 Ultra
- Self-hosted; infrastructure cost varies
- Fits in one request
Nemotron 3 Ultra has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Inkling
- $0.159
- Fits in one request
- Nemotron 3 Ultra
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
Nemotron 3 Ultra has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationInklingNemotron 3 Ultra
Context window
Maximum documented context; output-token limits may be lower.
Inkling
1M
Nemotron 3 Ultra
1M
API model ID
Inkling
Not sourced
Nemotron 3 Ultra
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Inkling
$0.374 per 1M cached input tokens
Nemotron 3 Ultra
No comparable hosted API rate
Documented inputs
Inkling
Not sourced
Nemotron 3 Ultra
Not sourced
Documented outputs
Inkling
Not sourced
Nemotron 3 Ultra
Not sourced
Provider availability
Inkling
Not sourced
Nemotron 3 Ultra
Not sourced
Reasoning profile
Inkling
Hybrid
Nemotron 3 Ultra
Reasoning
Weight access
Inkling
Open Weight
Nemotron 3 Ultra
Open Weight
License
Inkling
Open Weight
Nemotron 3 Ultra
Open Weight
Release date
Inkling
2026-07-15
Nemotron 3 Ultra
2026-06-04
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- Inkling has the higher public score, 67.02 versus 46.32, and the 90% score intervals do not overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Both models list 1M.
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence25 rows
Agentic
Terminal-Bench 2.0
Inkling63.8%SourceNemotron 3 Ultra56.4%SourceInkling leads this result
BrowseComp
Inkling77.1%SourceNemotron 3 Ultra44.4%SourceInkling leads this result
MCP Atlas
Inkling74.1%SourceNemotron 3 Ultra—Not directly comparable
PinchBench
Inkling—Nemotron 3 Ultra90.0%SourceNot directly comparable
τ³-bench results
Inkling—Nemotron 3 Ultra70.9%SourceNot directly comparable
HLE w/ tools
Inkling—Nemotron 3 Ultra37.4%SourceNot directly comparable
Coding
SWE-bench Verified
Inkling77.6%SourceNemotron 3 Ultra71.9%SourceInkling leads this result
SWE-bench Pro
Inkling54.3%SourceNemotron 3 Ultra—Not directly comparable
Terminal-Bench 2.0
Inkling63.8%SourceNemotron 3 Ultra56.4%SourceInkling leads this result
SWE Multilingual
Inkling—Nemotron 3 Ultra67.7%SourceNot directly comparable
LiveCodeBench v6
Inkling—Nemotron 3 Ultra89.0%SourceNot directly comparable
SciCode
Inkling—Nemotron 3 Ultra44.6%SourceNot directly comparable
Reasoning
Knowledge
GPQA
Inkling87.9%SourceNemotron 3 Ultra87%SourceInkling leads this result
GPQA-D
Inkling87.9%SourceNemotron 3 Ultra87.0%SourceInkling leads this result
HLE
Inkling46%SourceNemotron 3 Ultra26.7%SourceInkling leads this result
HLE w/o tools
Inkling30%SourceNemotron 3 Ultra26.7%SourceInkling leads this result
MMLU-Pro
Inkling—Nemotron 3 Ultra86.8%SourceNot directly comparable
Math
AIME26
Inkling97.1%SourceNemotron 3 Ultra—Not directly comparable
Multilingual
MMLU-ProX
Inkling—Nemotron 3 Ultra83%SourceNot directly comparable
Multimodal
Instruction following
Frequently asked questions
Which is better, Inkling or Nemotron 3 Ultra?
Inkling has the higher public score, 67.02 versus 46.32, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.
Which is better for coding, Inkling or Nemotron 3 Ultra?
The current coding averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, Inkling or Nemotron 3 Ultra?
Inkling leads the like-for-like agentic tasks comparison across 2 shared weighted benchmark rows.
Which costs less, Inkling or Nemotron 3 Ultra?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, Inkling or Nemotron 3 Ultra?
Both models list the same context window, 1M.
Related comparisons
Last updated August 29, 2026
Watch Inkling vs Nemotron 3 Ultra
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.