Skip to main content
BenchLM
Advertise

Model comparison

MiniMax M3 vs Nemotron 3 Ultra

Data verified

Head-to-head evidence from 25 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

MiniMax M3
MiniMax
69.75/100
No comparison
N/A
2 category wins0 category wins

Public leaderboard positions: MiniMax M3 #15 (Supported); Nemotron 3 Ultra unranked (Not scored). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. MiniMax M3 and Nemotron 3 Ultra share 25 comparable benchmark results. 2 of 8 categories are comparable. 20 results are unique to MiniMax M3; 13 to Nemotron 3 Ultra.

Updated July 23, 2026
Shared results
25
MiniMax M3 only
20
Nemotron 3 Ultra only
13
Comparable categories
2 / 8

Treat this as a split decision. MiniMax M3 makes more sense if agentic is the priority or you would rather avoid the extra latency and token burn of a reasoning model; Nemotron 3 Ultra is the better fit if you want the cheaper token bill or you want the stronger reasoning-first profile.

Confidence note. This is a partial-evidence comparison with 25 shared benchmark results across 6 evidence categories; 2 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

MiniMax M3 and Nemotron 3 Ultra finish on the same BenchAlign overall score, so this is less about a single winner and more about where the edge shows up. The BenchAlign headline says tie; the benchmark table is where the real choice happens.

MiniMax M3 is also the more expensive model on tokens at $0.30 input / $1.20 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Nemotron 3 Ultra. That is roughly Infinityx on output cost alone. Nemotron 3 Ultra is the reasoning model in the pair, while MiniMax M3 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for MiniMax M3 and Nemotron 3 Ultra
CategoryMiniMax M3ΔNemotron 3 Ultra
AgenticMiniMax M372.3Margin← 21.0Nemotron 3 Ultra51.3
CodingMiniMax M372.2Margin← 13.9Nemotron 3 Ultra58.3
ReasoningMiniMax M3Not measuredMarginNo overlapNemotron 3 Ultra61.9
KnowledgeMiniMax M3Not measuredMarginNo overlapNemotron 3 Ultra53.8
MathMiniMax M385.7MarginNo overlapNemotron 3 UltraNot measured
MultilingualMiniMax M3Not measuredMarginNo overlapNemotron 3 Ultra83.0
MultimodalMiniMax M364.9MarginNo overlapNemotron 3 UltraNot measured
Inst. FollowingMiniMax M3Not measuredMarginNo overlapNemotron 3 Ultra81.7

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · MiniMax M3B · Nemotron 3 Ultra

  1. BrowseComp

    Agentic
    Source ↗
    A 83.5%B 44.4%
    Winner: MiniMax M3Δ 39.1
    BrowseComp: MiniMax M3 scored 83.5%; Nemotron 3 Ultra scored 44.4%. MiniMax M3 wins this benchmark.

  2. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 66%B 56.4%
    Winner: MiniMax M3Δ 9.6
    Terminal-Bench 2.0: MiniMax M3 scored 66%; Nemotron 3 Ultra scored 56.4%. MiniMax M3 wins this benchmark.

  3. SWE-bench Verified

    Coding
    Source ↗
    A 80.5%B 71.9%
    Winner: MiniMax M3Δ 8.6
    SWE-bench Verified: MiniMax M3 scored 80.5%; Nemotron 3 Ultra scored 71.9%. MiniMax M3 wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricMiniMax M3Nemotron 3 UltraComparison
Input / output priceUSD per 1M tokensMiniMax M3$0.3 input / $1.2 outputNemotron 3 Ultra$0 input / $0 outputNemotron 3 Ultra has the lower combined listed price.
Generation speedtokens per secondMiniMax M3Not availableNemotron 3 UltraNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenMiniMax M3Not availableNemotron 3 UltraNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensMiniMax M31MNemotron 3 Ultra1MListed context windows are equal.

Benchmark Deep Dive

AgenticMiniMax M3 wins
BenchmarkMiniMax M3Nemotron 3 UltraResult
Terminal-Bench 2.0Source 66%56.4%MiniMax M3 leads
BrowseCompSource 83.5%44.4%MiniMax M3 leads
OSWorld-VerifiedSource 70.1%—Not comparable
MCP AtlasSource 74.2%—Not comparable
Claw-EvalSource 74.5%—Not comparable
AA Agentic IndexSource 35.4%27.4%MiniMax M3 leads
τ²-bench resultsSource 88.9%83.3%MiniMax M3 leads
GDPval-AASource 44.7%33.2%MiniMax M3 leads
GDPval-AASource 13951164MiniMax M3 leads
GDPval rubricsSource 74.7%—Not comparable
BankerToolBenchSource 76.1%—Not comparable
ResearchClawBenchSource 19.8%—Not comparable
OSWorld 2.0Source 4.6%—Not comparable
AA BriefcaseSource 1110870MiniMax M3 leads
AA EnterpriseOps-GymSource 32.1%28.9%MiniMax M3 leads
AA Harvey LABSource 88.4%81.7%MiniMax M3 leads
terminalBenchHardSource 42.4%36.4%MiniMax M3 leads
aaTerminalBench21Source 65.2%—Not comparable
PinchBenchSource —90.0%Not comparable
τ³-bench resultsSource —70.9%Not comparable
HLE w/ toolsSource —37.4%Not comparable
CodingMiniMax M3 wins
BenchmarkMiniMax M3Nemotron 3 UltraResult
SWE-bench VerifiedSource 80.5%71.9%MiniMax M3 leads
SWE-bench ProSource 59%—Not comparable
Terminal-Bench 2.0Source 66.0%56.4%MiniMax M3 leads
NL2RepoSource 42.1%—Not comparable
AA Coding IndexSource 58.6%49.3%MiniMax M3 leads
AA-SciCodeSource 45.4%39.9%MiniMax M3 leads
VIBE V2Source 50.1%—Not comparable
SVG-BenchSource 63.7%—Not comparable
KernelBench HardSource 28.8%—Not comparable
SWE MultilingualSource —67.7%Not comparable
SciCodeSource —44.6%Not comparable
Reasoning
BenchmarkMiniMax M3Nemotron 3 UltraResult
AA-LCRSource 74.0%67.0%MiniMax M3 leads
CritPtSource 3.7%3.1%MiniMax M3 leads
LongBench v2Source —61.9%Not comparable
Knowledge
BenchmarkMiniMax M3Nemotron 3 UltraResult
Artificial Analysis Intelligence IndexSource 44.4%37.8%MiniMax M3 leads
AA-GPQA DiamondSource 92.9%86.7%MiniMax M3 leads
AA-HLESource 37.1%26.6%MiniMax M3 leads
AA-Omniscience IndexSource 1.4%-0.8%MiniMax M3 leads
AA-Omniscience AccuracySource 15.0%21.6%Nemotron 3 Ultra leads
AA-Omniscience Hallucination RateSource 16.1%28.5%MiniMax M3 leads
AA Openness IndexSource 33.3%83.3%Nemotron 3 Ultra leads
GPQASource —87%Not comparable
GPQA-DSource —87.0%Not comparable
HLESource —26.7%Not comparable
HLE w/o toolsSource —26.7%Not comparable
MMLU-ProSource —86.8%Not comparable
Math
BenchmarkMiniMax M3Nemotron 3 UltraResult
USAMO 2026Source 85.7%—Not comparable
Multilingual
BenchmarkMiniMax M3Nemotron 3 UltraResult
MMLU-ProXSource —83%Not comparable
Multimodal
BenchmarkMiniMax M3Nemotron 3 UltraResult
OfficeQA ProSource 45.1%—Not comparable
OmniDocBench 1.5Source 91.6%—Not comparable
MMMU-ProSource 78.1%—Not comparable
VideoMMMUSource 84.6%—Not comparable
Video-MME (with subtitle)Source 85.4%—Not comparable
Design Arena WebsiteSource 12891129MiniMax M3 leads
AA-MMMU-ProSource 78.6%—Not comparable
Inst. Following
BenchmarkMiniMax M3Nemotron 3 UltraResult
AA-IFBenchSource 82.9%81.4%MiniMax M3 leads
IFBenchSource —81.7%Not comparable
Frequently Asked Questions (3)

Which is better, MiniMax M3 or Nemotron 3 Ultra?

MiniMax M3 and Nemotron 3 Ultra are tied on the BenchAlign overall score, so the right pick depends on which category matters most for your use case.

Which is better for coding, MiniMax M3 or Nemotron 3 Ultra?

MiniMax M3 has the edge for coding in this comparison, averaging 72.2 versus 58.3. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, MiniMax M3 or Nemotron 3 Ultra?

MiniMax M3 has the edge for agentic tasks in this comparison, averaging 72.3 versus 51.3. Inside this category, AA Briefcase is the benchmark that creates the most daylight between them.

Related Comparisons

Explore More

MiniMax Compare Pricing Methodology Find Your Best LLM Overall Rankings
Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.