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 | 7B/13B/70B |
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 | 4k |
| 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. | $0.256DeepInfra | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $0.712DeepInfra | — |
| $103StreamLake | — | |
| Benchmarks | ||
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. | 88.5% | 68.9% |
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. | 42% | — |
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. | 40.5% | — |
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. | — | 29.9% |
MATHReasoning — accuracy on the original competition mathematics test set | — | 13.5% |
ARC-ChallengeReasoning — multiple-choice science questions from the AI2 Reasoning Challenge; unrelated to ARC-AGI | — | 57.4% |
HellaSwagReasoning — commonsense completion of short scenarios | — | 85.3% |
WinoGrandeReasoning — resolving ambiguous pronouns in context | — | 80.2% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 59.1% | — |
GSM8KGrade-school math — Grade-school maths word problems that take a few steps of arithmetic to work through. It separated the models of 2022 and 2023 sharply, then saturated. One caveat on the historical numbers: OpenAI included part of the GSM8K training set in GPT-4's pre-training mix, so GPT-4's score is not a clean few-shot result. Higher is better. | — | 56.8% |
| Overview | ||
| Company | DeepSeek | Meta |
| Release date | Dec 26 2024 | Jul 18 2023 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
DeepSeek V3vsClaude Fable 5.2LLaMA 2vsClaude Fable 5.2DeepSeek V3vsGPT-6.2 AstraLLaMA 2vsGPT-6.2 AstraDeepSeek V3vsGemini 4 HydrogenLLaMA 2vsGemini 4 HydrogenDeepSeek V3vsGrok 4.7LLaMA 2vsGrok 4.7DeepSeek V3vsMistral Large 4LLaMA 2vsMistral Large 4DeepSeek V3vsKimi K3.1 ProLLaMA 2vsKimi K3.1 ProFrequently asked questions
DeepSeek V3 leads LLaMA 2 on 1 of the 1 benchmark they both report (MMLU). LLaMA 2 shipped 527 days before DeepSeek V3, so benchmark comparisons should account for the intervening progress.
DeepSeek V3 has 671B parameters, while LLaMA 2 has 7B/13B/70B. Context windows are 128k (DeepSeek V3) vs 4k (LLaMA 2).
On MMLU, DeepSeek V3 leads at 88.5% vs LLaMA 2 at 68.9%.