Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

See the free Radar Brief
BenchLM
Microsoft logo
Model A
MAI-Thinking-1

Microsoft

52.27/100

Estimated · Public rank #123

90% interval 42.4–62.1

MAI-Thinking-1 vs Nemotron 3 Ultra

Updated September 4, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

NVIDIA logo
Model B
Nemotron 3 Ultra

NVIDIA

40.67/100

Estimated · Public rank #188

90% interval 30.8–50.5

Decision reading

MAI-Thinking-1 has the higher public score estimate, 52.27 versus 40.67, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.

8 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Share or export

Share on XLinkedInSocial cardCSVJSON

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.

  • Long documents

    Prompts that approach the documented context limit

    Nemotron 3 Ultra

    Nemotron 3 Ultra has the larger documented context window.

    Confidence: documented

Show secondary and unsupported calls

  • Coding work

    Code generation, repair, and software-engineering tasks

    Not enough matched evidence

    MAI-Thinking-1 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.

    Confidence: limited

  • Agentic work

    Tool use, computer use, and multi-step task completion

    Not enough matched evidence

    MAI-Thinking-1 is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.

    Confidence: limited

  • 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.

Shared results
8
MAI-Thinking-1 only
6
Nemotron 3 Ultra only
16
Like-for-like categories
1 / 8

3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.

Instruction following

Like-for-like
MAI-Thinking-1
94.7
#1/120
Nemotron 3 Ultra
88.7
#28/120
Basis
Provisional lane · 1 vs 1 weighted rows
Reading
MAI-Thinking-1 leads

Agentic

Directional only
MAI-Thinking-1
51.7
Estimated · #53/151
Nemotron 3 Ultra
22.2
Supported · #150/151
Basis
BenchAlign lane · 1 vs 6 public rows
Reading
Directional only

Coding

Directional only
MAI-Thinking-1
51.8
Estimated · #60/183
Nemotron 3 Ultra
26.9
Supported · #174/183
Basis
BenchAlign lane · 4 vs 7 public rows
Reading
Directional only

Knowledge

Directional only
MAI-Thinking-1
53.3
Estimated · #68/181
Nemotron 3 Ultra
45.6
Estimated · #114/181
Basis
BenchAlign lane · 4 vs 7 public rows
Reading
Directional only

Reasoning

Not comparable
MAI-Thinking-1
Not ranked
Nemotron 3 Ultra
48.7
Unranked · 1 rankable row
Basis
Provisional lane · 0 vs 1 weighted rows
Reading
Not comparable

Math

Not comparable
MAI-Thinking-1
73.4
Unranked · 3 rankable rows
Nemotron 3 Ultra
Not ranked
Basis
Provisional lane · 2 vs 0 weighted rows
Reading
Not comparable

Multilingual

Not comparable
MAI-Thinking-1
Not ranked
Nemotron 3 Ultra
47.4
#7/12
Basis
Provisional lane · 0 vs 1 weighted rows
Reading
Not comparable

Multimodal

Not comparable
MAI-Thinking-1
Not ranked
Nemotron 3 Ultra
Not ranked
Basis
Provisional lane · 0 vs 0 weighted rows
Reading
Not comparable

Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.

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.

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

MAI-Thinking-1
API rate not published
Fits in one request
Nemotron 3 Ultra
Self-hosted; infrastructure cost varies
Fits in one request

MAI-Thinking-1 has no comparable published API token rate. Nemotron 3 Ultra has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

MAI-Thinking-1
API rate not published
Fits in one request
Nemotron 3 Ultra
Self-hosted; infrastructure cost varies
Fits in one request

MAI-Thinking-1 has no comparable published API token rate. Nemotron 3 Ultra has no comparable published API token rate.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

MAI-Thinking-1
API rate not published
Fits in one request
Cached-input rate unavailable
Nemotron 3 Ultra
Self-hosted; infrastructure cost varies
Fits in one request
Cached-input rate unavailable

MAI-Thinking-1 has no comparable published API token rate. 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.

Context window

Maximum documented context; output-token limits may be lower.

MAI-Thinking-1

256K

Nemotron 3 Ultra

1M

API model ID

MAI-Thinking-1

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.

MAI-Thinking-1

No comparable hosted API rate

Nemotron 3 Ultra

No comparable hosted API rate

Documented inputs

MAI-Thinking-1

Not sourced

Nemotron 3 Ultra

Not sourced

Documented outputs

MAI-Thinking-1

Not sourced

Nemotron 3 Ultra

Not sourced

Provider availability

MAI-Thinking-1

Not sourced

Nemotron 3 Ultra

Not sourced

Reasoning profile

MAI-Thinking-1

Reasoning

Nemotron 3 Ultra

Reasoning

Weight access

MAI-Thinking-1

Proprietary

Nemotron 3 Ultra

Open Weight

License

MAI-Thinking-1

Proprietary

Nemotron 3 Ultra

Open Weight

Release date

MAI-Thinking-1

2026-06-02

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
MAI-Thinking-1 has the higher public score estimate, 52.27 versus 40.67, but the 90% score intervals overlap.
Workload cost
A complete comparable API-rate estimate is not available for both models.
Context tradeoff
Nemotron 3 Ultra has the larger documented window (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 evidence30 rows

Agentic

  • Terminal-Bench 2.0

    MAI-Thinking-146%
    Source
    Nemotron 3 Ultra56.4%
    Source

    Nemotron 3 Ultra leads this result

  • PinchBench

    MAI-Thinking-1—
    Nemotron 3 Ultra90.0%
    Source

    Not directly comparable

  • BrowseComp

    MAI-Thinking-1—
    Nemotron 3 Ultra44.4%
    Source

    Not directly comparable

  • τ³-bench results

    MAI-Thinking-1—
    Nemotron 3 Ultra70.9%
    Source

    Not directly comparable

  • HLE w/ tools

    MAI-Thinking-1—
    Nemotron 3 Ultra37.4%
    Source

    Not directly comparable

  • Terminal-Bench 2.1 (Vals)

    MAI-Thinking-1—
    Nemotron 3 Ultra50.9%
    Source

    Not directly comparable

Coding

  • LiveCodeBench v6

    MAI-Thinking-187.7%
    Source
    Nemotron 3 Ultra89.0%
    Source

    Nemotron 3 Ultra leads this result

  • SWE-bench Verified

    MAI-Thinking-173.5%
    Source
    Nemotron 3 Ultra71.9%
    Source

    MAI-Thinking-1 leads this result

  • SWE-bench Pro

    MAI-Thinking-152.8%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

  • Terminal-Bench 2.0

    MAI-Thinking-146.0%
    Source
    Nemotron 3 Ultra56.4%
    Source

    Nemotron 3 Ultra leads this result

  • SWE Multilingual

    MAI-Thinking-1—
    Nemotron 3 Ultra67.7%
    Source

    Not directly comparable

  • SciCode

    MAI-Thinking-1—
    Nemotron 3 Ultra44.6%
    Source

    Not directly comparable

  • LiveCodeBench (Vals)

    MAI-Thinking-1—
    Nemotron 3 Ultra86.0%
    Source

    Not directly comparable

  • SWE-bench (Vals)

    MAI-Thinking-1—
    Nemotron 3 Ultra69.0%
    Source

    Not directly comparable

Reasoning

  • Graphwalks BFS 128K

    MAI-Thinking-190%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

  • CritPt

    MAI-Thinking-1—
    Nemotron 3 Ultra3.1%
    Source

    Not directly comparable

  • LongBench v2

    MAI-Thinking-1—
    Nemotron 3 Ultra61.9%
    Source

    Not directly comparable

Knowledge

  • GPQA

    MAI-Thinking-184.2%
    Source
    Nemotron 3 Ultra87%
    Source

    Nemotron 3 Ultra leads this result

  • GPQA-D

    MAI-Thinking-184.2%
    Source
    Nemotron 3 Ultra87.0%
    Source

    Nemotron 3 Ultra leads this result

  • MMLU-Pro

    MAI-Thinking-185%
    Source
    Nemotron 3 Ultra86.8%
    Source

    Nemotron 3 Ultra leads this result

  • SimpleQA

    MAI-Thinking-131%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

  • HLE

    MAI-Thinking-1—
    Nemotron 3 Ultra26.7%
    Source

    Not directly comparable

  • HLE w/o tools

    MAI-Thinking-1—
    Nemotron 3 Ultra26.7%
    Source

    Not directly comparable

  • GPQA Diamond (Vals)

    MAI-Thinking-1—
    Nemotron 3 Ultra86.1%
    Source

    Not directly comparable

  • MMLU-Pro (Vals)

    MAI-Thinking-1—
    Nemotron 3 Ultra85.8%
    Source

    Not directly comparable

Math

  • AIME 2025

    MAI-Thinking-197%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

  • AIME26

    MAI-Thinking-194.5%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

  • HMMT Feb 2026

    MAI-Thinking-184.9%
    Source
    Nemotron 3 Ultra—

    Not directly comparable

Multilingual

  • MMLU-ProX

    MAI-Thinking-1—
    Nemotron 3 Ultra83%
    Source

    Not directly comparable

Instruction following

  • IFBench

    MAI-Thinking-185%
    Source
    Nemotron 3 Ultra81.7%
    Source

    MAI-Thinking-1 leads this result

Frequently asked questions

Which is better, MAI-Thinking-1 or Nemotron 3 Ultra?

MAI-Thinking-1 has the higher public score estimate, 52.27 versus 40.67, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.

Which is better for coding, MAI-Thinking-1 or Nemotron 3 Ultra?

MAI-Thinking-1 scores higher for coding on the public lane, 51.8 to 26.9. MAI-Thinking-1 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.

Which is better for agentic tasks, MAI-Thinking-1 or Nemotron 3 Ultra?

MAI-Thinking-1 scores higher for agentic tasks on the public lane, 51.7 to 22.2. MAI-Thinking-1 is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.

Which costs less, MAI-Thinking-1 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, MAI-Thinking-1 or Nemotron 3 Ultra?

Nemotron 3 Ultra has the larger documented context window: 1M, compared with 256K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated September 4, 2026

Watch MAI-Thinking-1 vs Nemotron 3 Ultra

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 2,000+ readers.