DeepSeek V4 Flash 0731 vs NVIDIA Nemotron Nano 9B V2: Release Comparison

Comparison of the DeepSeek V4 Flash 0731 and NVIDIA Nemotron Nano 9B V2 releases: 2 and 2 models respectively, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across all 4 models.

  • For intelligence, DeepSeek V4 Flash 0731 scores highest: DeepSeek V4 Flash 0731 (Reasoning, Max Effort) at 52, against NVIDIA Nemotron Nano 9B V2 (Reasoning) at 9 for NVIDIA Nemotron Nano 9B V2.
  • For output speed, NVIDIA Nemotron Nano 9B V2 is fastest: NVIDIA Nemotron Nano 9B V2 (Reasoning) at 203 t/s, against DeepSeek V4 Flash Vision (Reasoning, Max Effort) at 109 t/s for DeepSeek V4 Flash 0731.

Releases compared

DeepSeekDeepSeek V4 Flash 0731NVIDIANVIDIA Nemotron Nano 9B V2

Intelligence

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 饾湉鲁-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Estimate (independent evaluation forthcoming)

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 饾湉鲁-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Intelligence Index vs. Cost per Intelligence Index Task

Artificial Analysis Intelligence Index 路 Weighted average cost (USD) per Artificial Analysis Intelligence Index task
Most attractive quadrant
Pareto line

Weighted average cost per Intelligence Index task. Each evaluation鈥檚 cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 饾湉鲁-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Cost

Cost per Intelligence Index Task

Weighted average cost (USD) per Artificial Analysis Intelligence Index task, segmented by token type. Lower is better

Weighted average cost per Intelligence Index task. Each evaluation鈥檚 cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Speed & Latency

Output Speed

Output tokens per second 路 Higher is better

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).

Capability Scores

Capability Indices

Measures the performance of models on specific capabilities and industries
Agentic Index

Measures performance in agentic workflows, focusing on behaviors like tool use, planning, autonomy, and complex problem solving.

Finance & Accounting Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, Humanity's Last Exam, 饾湉鲁-Banking, AA-LCR 路 Higher is better

Strategy & Ops Index

Incorporates 4 evaluations: AA-Omniscience, GDPval-AA v2, 饾湉鲁-Banking, AA-LCR 路 Higher is better

Legal Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, AA-LCR, 饾湉鲁-Banking, Humanity's Last Exam 路 Higher is better

Healthcare & Medical Index

Incorporates 5 evaluations: AA-Omniscience, GDPval-AA v2, MLCR-AA, Humanity's Last Exam, 饾湉鲁-Banking 路 Higher is better

Engineering Index

Incorporates 6 evaluations: AA-Omniscience, Humanity's Last Exam, GPQA Diamond, CritPt, GDPval-AA v2, Terminal-Bench v2.1 路 Higher is better

Economics Index

Incorporates 4 evaluations: AA-Omniscience, Humanity's Last Exam, GDPval-AA v2, AA-LCR 路 Higher is better

Further details

Weights
Provider Benchmarks
DeepSeek V4 Flash 0731 (Reasoning, Max Effort)
DeepSeek logoDeepSeek
52
284B
13B active at inference time
1M
$0.2
108
馃
CoreWeaveWaferNebius
+11
View
DeepSeek V4 Flash Vision (Reasoning, Max Effort)
DeepSeek logoDeepSeek
51
284B
13B active at inference time
1M
$0.2
109
Not available
DeepSeek
View
NVIDIA Nemotron Nano 9B V2 (Reasoning)
NVIDIA logoNVIDIA
9
9B
131k
$0.1
203
馃
DeepInfra
View
NVIDIA Nemotron Nano 9B V2 (Non-reasoning)
NVIDIA logoNVIDIA
7
9B
131k
$0.1
170
馃
DeepInfraAmazon Bedrock
View