What are Llama Nemotron Ultra 253B's benchmark scores?

As of 25 Oct 2026, Llama Nemotron Ultra 253B scores 48.1 ±9.9 on the BenchLeader Index, strongest in maths (55) and weakest in agents & tools (24). That rests on 5 independent benchmark results, and every score on this page links to its publisher.

How does Llama Nemotron Ultra 253B rank?

Llama Nemotron Ultra 253B is #438 of 785 ranked configurations on 25 Oct 2026. 557 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.

When was Llama Nemotron Ultra 253B released?

NVIDIA released Llama Nemotron Ultra 253B on 7 Apr 2025. BenchLeader has results for it at 2 reasoning-effort settings.

NVIDIAOpen weightsReasoning modelCompare with another model

Llama Nemotron Ultra 253B

Reasoning effort

Llama Nemotron Ultra 253B is a NVIDIA open-weights reasoning model, released 7 Apr 2025. Its best configuration ranks #438 of 785 on the BenchLeader Index at 48.1 ±9.9, in the lower half. The ± is the point: 557 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in maths (55) and lowest in agents & tools (24). It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.

Blended price: –
Output speed: –
First answer: –
Context: 128k
Full answer: –
Cost per run: –
Released: 7 Apr 2025

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index48
Reasoning53
Coding51
Agents & tools24
Maths55
Human preference51

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. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
thinking	43.3	#573	–	–	–	Agents & tools 36 · Composite 38 · Instruction following 41 · Knowledge 41 · Reasoning 41
not statedbest	48.1	#438	–	–	–	Agents & tools 24 · Coding 51 · Human preference 51 · Maths 55 · Reasoning 53

How its index has moved

19 Sept 2026 to 23 Oct 2026

not stated
thinking

The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.

Benchmark results

See all LLM benchmarks. 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	not stated	Source
Arena Hard Promptseffort not stated	–	1375#192	Arena (LMArena)
GPQA Diamond (AA)not in index	72.8%#273	–	Artificial Analysis
Humanity's Last Exam (AA)not in index	7.4%#367	–	Artificial Analysis
CritPt	0.0%#315	–	Artificial Analysis

Coding

Benchmark	not stated	Source	Trend
Arena Coding	1391#201	Arena (LMArena)

Agents & tools

Benchmark	thinking	not stated	Source
Terminal-Bench Hard	2.3%#320	–	Artificial Analysis
τ²-Bench Telecom (AA)not in index	11.4%#376	–	Artificial Analysis
BFCL Overalleffort not stated	–	10.0%#82	Berkeley Function Calling Leaderboard

Maths

Benchmark	not stated	Source	Trend
Arena Maths	1381#172	Arena (LMArena)

Knowledge

Benchmark	thinking	Source
AA-Omniscience	-44.9#351	Artificial Analysis
AA-Omniscience: accuracynot in index	20.1%#306	Artificial Analysis
AA-Omniscience: non-hallucinationnot in index	18.8%#295	Artificial Analysis

Instruction following

Benchmark	thinking	not stated	Source	Trend
IFBench	38.2%#281	–	Artificial Analysis
Arena Instruction Followingnot in indexeffort not stated	–	1351#188	Arena (LMArena)

Human preference

Benchmark	not stated	Source
Arena Text	1348#204	Arena (LMArena)
Arena Creative Writingnot in index	1331#188	Arena (LMArena)
Arena Multi-turnnot in index	1342#204	Arena (LMArena)
Arena Longer Queriesnot in index	1342#212	Arena (LMArena)

Composite

Benchmark	thinking	Source	Trend
AA Intelligence Index v4.3.2	7.5#469	Artificial Analysis

New this month

Questions people ask

What are Llama Nemotron Ultra 253B's benchmark scores?: As of 25 Oct 2026, Llama Nemotron Ultra 253B scores 48.1 ±9.9 on the BenchLeader Index, strongest in maths (55) and weakest in agents & tools (24). That rests on 5 independent benchmark results, and every score on this page links to its publisher.
How does Llama Nemotron Ultra 253B rank?: Llama Nemotron Ultra 253B is #438 of 785 ranked configurations on 25 Oct 2026. 557 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
When was Llama Nemotron Ultra 253B released?: NVIDIA released Llama Nemotron Ultra 253B on 7 Apr 2025. BenchLeader has results for it at 2 reasoning-effort settings.

Data as of 25 Oct 2026. Compare these configurations.

Cite as: BenchLeader, “Llama Nemotron Ultra 253B: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/nemotron-ultra-253b.