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. | 284B | 21B |
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. | 1M | 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. | $0.0434Wafer | $0.018Darkbloom |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $0.14Relace | $0.068PrimeIntellect |
| $21.00Relace · fp4 | $8.10Darkbloom · fp8 | |
| Benchmarks | ||
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. | 79% | 60.7% |
Humanity's Last Exam · with toolsMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | 34.8% | 17.3% |
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. | 39% | — |
Terminal-Bench 2.1Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | 82.7% | — |
Toolathlon-VerifiedPersonal tool use — Tests how well the AI uses everyday personal tools and apps to get things done — a human-checked version of Toolathlon. Higher is better. | 70.3% | — |
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better. | 73.2% | — |
CyberGymCybersecurity — Tests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better. | 76.7% | — |
Humanity's Last Exam · no toolsMultidisciplinary reasoning — Humanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better. | — | 10.9% |
ARC-AGI-2Abstract reasoning — Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better. | 61.4% | — |
AIME 2024Reasoning — single-answer accuracy on the 2024 invitational mathematics exam without tools | — | 92.1% |
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | — | 91.7% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | — | 71.5% |
MMMLUKnowledge — average accuracy on multilingual translations of MMLU | — | 75.7% |
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. | — | 85.3% |
AutomationBenchBusiness workflows — Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better. | 25.1% | — |
HealthBench · full setHealthcare — physician rubric score on the full set of the original HealthBench evaluation | — | 42.5% |
HealthBench · hardHealthcare — physician rubric score on the hard subset of the original HealthBench evaluation | — | 10.8% |
HealthBench · consensusHealthcare — physician rubric score on the consensus subset of the original HealthBench evaluation | — | 82.6% |
| Overview | ||
| Company | DeepSeek | OpenAI |
| Release date | Jul 31 2026 | Aug 5 2025 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
DeepSeek-V4-Flash-0731vsClaude Fable 5.2gpt-oss-20bvsClaude Fable 5.2DeepSeek-V4-Flash-0731vsGemini 4 Carbongpt-oss-20bvsGemini 4 CarbonDeepSeek-V4-Flash-0731vsMuse Embergpt-oss-20bvsMuse EmberDeepSeek-V4-Flash-0731vsGrok 4.7gpt-oss-20bvsGrok 4.7DeepSeek-V4-Flash-0731vsMistral Large 4gpt-oss-20bvsMistral Large 4DeepSeek-V4-Flash-0731vsKimi K3.1 Progpt-oss-20bvsKimi K3.1 ProFrequently asked questions
DeepSeek-V4-Flash-0731 leads gpt-oss-20b on 2 of the 2 benchmarks they both report (SWE-Bench Verified, Humanity's Last Exam). gpt-oss-20b shipped 360 days before DeepSeek-V4-Flash-0731, so benchmark comparisons should account for the intervening progress.
DeepSeek-V4-Flash-0731 has 284B parameters, while gpt-oss-20b has 21B. Context windows are 1M (DeepSeek-V4-Flash-0731) vs 128k (gpt-oss-20b).
On SWE-Bench Verified, DeepSeek-V4-Flash-0731 leads at 79% vs gpt-oss-20b at 60.7%. On Humanity's Last Exam · with tools, DeepSeek-V4-Flash-0731 leads at 34.8% vs gpt-oss-20b at 17.3%.