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/14B |
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 | 1M |
| 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. | $0.259SiliconFlow | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $0.38DeepInfra | — |
| $71.00DeepInfra · fp4 | — | |
| Benchmarks | ||
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. | 83.3% | 38.6% |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 85% | 63.3% |
BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | 13% | — |
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. | 73.1% | — |
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. | 70.2% | — |
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. | — | 88.4% |
Terminal-Bench 2.0Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? (Version 2.0 of the test.) Higher is better. | 46.4% | — |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 84.3% |
ToolathlonGeneral tool use — Tests how well the AI uses everyday real-world tools and apps to get things done. Higher is better. | 35.2% | — |
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better. | 51.4% | — |
Humanity's Last Exam · text only, no toolsMultidisciplinary reasoning — accuracy on the 2,158 text-only questions without tools | 25.1% | — |
MATHReasoning — accuracy on the original competition mathematics test set | — | 79.5% |
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | 93.1% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 82.4% | — |
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. | — | 94.8% |
| Overview | ||
| Company | DeepSeek | Qwen |
| Release date | Dec 1 2025 | Jan 27 2025 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
DeepSeek-V3.2vsClaude Fable 5.2Qwen2.5-1MvsClaude Fable 5.2DeepSeek-V3.2vsGPT-6.2 AstraQwen2.5-1MvsGPT-6.2 AstraDeepSeek-V3.2vsGemini 4 NeonQwen2.5-1MvsGemini 4 NeonDeepSeek-V3.2vsMuse EmberQwen2.5-1MvsMuse EmberDeepSeek-V3.2vsGrok 4.7Qwen2.5-1MvsGrok 4.7DeepSeek-V3.2vsMistral Large 4Qwen2.5-1MvsMistral Large 4Frequently asked questions
DeepSeek-V3.2 leads Qwen2.5-1M on 2 of the 2 benchmarks they both report (LiveCodeBench, MMLU-Pro). Qwen2.5-1M shipped 308 days before DeepSeek-V3.2, so benchmark comparisons should account for the intervening progress.
DeepSeek-V3.2 has 671B parameters, while Qwen2.5-1M has 7B/14B. Context windows are 128k (DeepSeek-V3.2) vs 1M (Qwen2.5-1M).
On LiveCodeBench, DeepSeek-V3.2 leads at 83.3% vs Qwen2.5-1M at 38.6%. On MMLU-Pro, DeepSeek-V3.2 leads at 85% vs Qwen2.5-1M at 63.3%.