modelbenchmark.io

Nemotron 3 Super 120B A12B

hf

Compare

Nemotron middle tier for collaborative agents and high-volume reasoning workloads

Specification

most-agreed values

Context
262K
Max output
66K
Released
2026-03-11
Knowledge cutoff
2025-06
Retires
—
Open weights
yes
Input
text
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.30
Output
$1.00
Cache read
$0.06

Quality

9 benchmarks · 2 sources

Composite
78th
percentile of 344 scored models
Rank
77 / 344
±0.08 sd
Evidence
9 × 2
3 or more benchmarks, from 2 or more sources
Effort range
16.5
points between effort settings
reasoning76th3/4 bench
math85th3/7 bench
knowledge70th1/1 bench
general77th1/1 bench

By reasoning effort

every setting placed on the same scale as the leaderboard

SettingCompositeBenchmarks behind it
high89th3
max81st8
low57th4

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
GPQA diamond93.1 ±1.5maxEpoch AI2026-07-16
↳91.9 ±1.9highEpoch AI2026-08-07
↳84.8 ±2.6lowEpoch AI2026-08-07
Settings differ by 8.3 points on this benchmark.
OTIS Mock AIME 2024-202597.2 ±1.1maxEpoch AI2026-07-16
↳93.3 ±3.8highEpoch AI2026-08-07
↳68.9 ±7.0lowEpoch AI2026-08-07
Settings differ by 28.3 points on this benchmark.
FrontierMath-Tiers-1-3-v272.2 ±2.7maxEpoch AI2026-07-17
LiveBench79.5default · 23 runsLiveBench2026-06-25
SimpleQA Verified50.6 ±1.6maxEpoch AI2026-08-27
Chess Puzzles37.0 ±4.9highEpoch AI2026-10-13
↳36.0 ±4.8maxEpoch AI2026-10-13
↳18.0 ±3.9lowEpoch AI2026-10-13
Settings differ by 19.0 points on this benchmark.
FrontierMath-Tier-4-v239.0 ±7.7maxEpoch AI2026-07-17
Mystery Game Puzzles26.0 ±4.4maxEpoch AI2026-07-29
↳12.0 ±3.3lowEpoch AI2026-08-29
Settings differ by 14.0 points on this benchmark.
Furniture Assembly34.2 ±6.1maxEpoch AI2026-09-11

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 12 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
synthetic · openai/gpt-oss-120b$0.10$0.10$0.02—131K66K—
synthetic · zai-org/GLM-4.7-Flash$0.10$0.50$0.02—197K66K—
synthetic · zai-org/GLM-5.3-Flash$0.15$0.50$0.04—524K66K—
synthetic · nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4$0.30$1.00$0.06—262K66K—
synthetic · Qwen/Qwen3.6-27B$0.45$3.60$0.45—262K66K—
synthetic · deepseek-ai/DeepSeek-V4.1-Flash$0.60$1.20$0.03—524K66K—
synthetic · deepseek-ai/DeepSeek-V4.1-Flash-1017$0.60$1.20$0.03—524K66K—
synthetic · MiniMaxAI/MiniMax-M3$0.60$1.20$0.60—524K66K—
synthetic · moonshotai/Kimi-K2.7-Code$0.95$4.00$0.95—262K66K—
synthetic · zai-org/GLM-5.2$1.40$4.40$1.40—524K66K—
synthetic · zai-org/GLM-5.3$1.40$4.40$0.26—524K66K—
synthetic · moonshotai/Kimi-K3$3.00$15.00$0.45—524K66K—

Price history

append-only observations · a listing writes a row only when its price moves

DateHostField$/M
2026-09-02synthetic · nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4Cache read$0.30
2026-10-22synthetic · nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4Cache read$0.06
2026-09-02synthetic · openai/gpt-oss-120bCache read$0.10
2026-10-22synthetic · openai/gpt-oss-120bCache read$0.02
2026-09-02synthetic · zai-org/GLM-4.7-FlashCache read$0.10
2026-10-22synthetic · zai-org/GLM-4.7-FlashCache read$0.02

Other listings of this model

this page is one of them

The corpus files Nemotron 3 Super 120B A12B under several keys. This page is the hf listing. The full record — every host, every price and every benchmark score — is on the main Nemotron 3 Super 120B A12B page.