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. | — | 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. | 200k | 256k |
| Benchmarks | ||
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 65% | 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. | 49% | — |
HumanEvalFunction synthesis — 164 small Python problems: the AI is given a function's description and has to write the working function. This was the coding benchmark of the GPT-3.5 and GPT-4 era, before the field moved to fixing real bugs in real repositories. Higher is better. | 93.7% | — |
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% |
MATH-500Reasoning — accuracy on the 500-problem MATH subset | — | 95.4% |
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | — | 78.5% |
MMLUGeneral knowledge — A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better. | 90.5% | — |
MGSMMultilingual math — solving grade-school word problems translated into multiple languages | 92.5% | — |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 70.4% | — |
| Overview | ||
| Company | Anthropic | NVIDIA |
| Release date | Oct 22 2024 | Mar 16 2026 |
| Access | Closed | Open Weight |
| Model details | View model | View model |
Other comparisons
Claude 3.5 Sonnet (upgraded)vsGPT-6.2 AstraNemotron 3 Nano 4BvsGPT-6.2 AstraClaude 3.5 Sonnet (upgraded)vsGemini 4 NeonNemotron 3 Nano 4BvsGemini 4 NeonClaude 3.5 Sonnet (upgraded)vsMuse EmberNemotron 3 Nano 4BvsMuse EmberClaude 3.5 Sonnet (upgraded)vsGrok 4.7Nemotron 3 Nano 4BvsGrok 4.7Claude 3.5 Sonnet (upgraded)vsDeepSeek-V4.1-Flash-1017Nemotron 3 Nano 4BvsDeepSeek-V4.1-Flash-1017Claude 3.5 Sonnet (upgraded)vsMistral Large 4Nemotron 3 Nano 4BvsMistral Large 4Frequently asked questions
Claude 3.5 Sonnet (upgraded) leads Nemotron 3 Nano 4B on 1 of the 1 benchmark they both report (GPQA Diamond). Claude 3.5 Sonnet (upgraded) shipped 510 days before Nemotron 3 Nano 4B, so benchmark comparisons should account for the intervening progress.
Context windows are 200k (Claude 3.5 Sonnet (upgraded)) vs 256k (Nemotron 3 Nano 4B). Claude 3.5 Sonnet (upgraded) is closed, while Nemotron 3 Nano 4B is open weight.
On GPQA Diamond, Claude 3.5 Sonnet (upgraded) leads at 65% vs Nemotron 3 Nano 4B at 53.2%.