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 | 1B/4B/12B/27B |
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 | 128k |
| 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. | $0.27SiliconFlow | $0.04DeepInfra |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $1.00SiliconFlow | $0.05Parasail |
| $104SiliconFlow · fp8 | $10.50Parasail · bf16 | |
| 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. | 74.9% | 29.7% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 80.7% | 42.4% |
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. | — | 3% |
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. | 68.4% | — |
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. | 57.8% | — |
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. | — | 87.8% |
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. | — | 95.9% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | — | 64.9% |
| Overview | ||
| Company | DeepSeek | |
| Release date | Sep 22 2025 | Mar 12 2025 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
DeepSeek V3.1 TerminusvsClaude Haiku 5.5Gemma 3vsClaude Haiku 5.5DeepSeek V3.1 TerminusvsGPT-6.1 SolGemma 3vsGPT-6.1 SolDeepSeek V3.1 TerminusvsMuse Spark OpenGemma 3vsMuse Spark OpenDeepSeek V3.1 TerminusvsGrok 4.7Gemma 3vsGrok 4.7DeepSeek V3.1 TerminusvsMistral Large 4Gemma 3vsMistral Large 4DeepSeek V3.1 TerminusvsKimi K3Gemma 3vsKimi K3Frequently asked questions
DeepSeek V3.1 Terminus leads Gemma 3 on 2 of the 2 benchmarks they both report (LiveCodeBench, GPQA Diamond). Gemma 3 shipped 194 days before DeepSeek V3.1 Terminus, so benchmark comparisons should account for the intervening progress.
DeepSeek V3.1 Terminus has 671B parameters, while Gemma 3 has 1B/4B/12B/27B. Context windows are 128k (DeepSeek V3.1 Terminus) vs 128k (Gemma 3).
On LiveCodeBench, DeepSeek V3.1 Terminus leads at 74.9% vs Gemma 3 at 29.7%. On GPQA Diamond, DeepSeek V3.1 Terminus leads at 80.7% vs Gemma 3 at 42.4%.