Compare AI models
| Specifications | ||
ParametersThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. | 671B | 4B |
Context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider. | 128k | 256k |
| API pricingUSD per 1M tokens · lower wins | ||
Cheapest inputLowest input rate across every provider on OpenRouter, the lab included. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $0.25DeepInfra | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $0.95DeepInfra | — |
| $97.50DeepInfra · fp4 | — | |
| Benchmarks | ||
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | 88.4% | 78.5% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 80.1% | 53.2% |
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 66% | — |
SWE-Bench MultilingualMultilingual coding — Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | 54.5% | — |
LiveCodeBenchCompetitive coding — Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better. | 74.8% | — |
τ²-bench · retailAgentic tasks — task success in the retail domain of the second-generation customer-service evaluation | — | 39.8% |
τ²-bench · airlineAgentic tasks — task success in the airline domain of the second-generation customer-service evaluation | — | 33.3% |
τ²-bench · telecomAgentic tasks — task success in the telecom domain of the second-generation customer-service evaluation | — | 33% |
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better. | 30% | — |
Humanity's Last Exam · text only, no toolsMultidisciplinary reasoning — accuracy on the 2,158 text-only questions without tools | 15.9% | — |
MATH-500Reasoning — accuracy on the 500-problem MATH subset | — | 95.4% |
AIME 2024Reasoning — single-answer accuracy on the 2024 invitational mathematics exam without tools | 93.1% | — |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 84.8% | — |
| Overview | ||
| Company | DeepSeek | NVIDIA |
| Release date | Aug 21 2025 | Mar 16 2026 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
DeepSeek V3.1vsClaude Fable 5.2Nemotron 3 Nano 4BvsClaude Fable 5.2DeepSeek V3.1vsGPT-6.2 AstraNemotron 3 Nano 4BvsGPT-6.2 AstraDeepSeek V3.1vsGemini 4 NeonNemotron 3 Nano 4BvsGemini 4 NeonDeepSeek V3.1vsMuse EmberNemotron 3 Nano 4BvsMuse EmberDeepSeek V3.1vsGrok 4.7Nemotron 3 Nano 4BvsGrok 4.7DeepSeek V3.1vsMistral Large 4Nemotron 3 Nano 4BvsMistral Large 4Frequently asked questions
DeepSeek V3.1 leads Nemotron 3 Nano 4B on 2 of the 2 benchmarks they both report (AIME 2025, GPQA Diamond). DeepSeek V3.1 shipped 207 days before Nemotron 3 Nano 4B, so benchmark comparisons should account for the intervening progress.
DeepSeek V3.1 has 671B parameters, while Nemotron 3 Nano 4B has 4B. Context windows are 128k (DeepSeek V3.1) vs 256k (Nemotron 3 Nano 4B).
On AIME 2025, DeepSeek V3.1 leads at 88.4% vs Nemotron 3 Nano 4B at 78.5%. On GPQA Diamond, DeepSeek V3.1 leads at 80.1% vs Nemotron 3 Nano 4B at 53.2%.