DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context...
Tested on 7 benchmarks · BenchGecko score 51.8. Top scores: Chatbot Arena Elo — Overall (1417.2%), Lech Mazur Writing (85.2%), Dtbench (71.1%).
MiniMax M2.7 scores 51.2 (99% as good) at $0.21/1M input · 16% cheaper
Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.
Deceptively simple questions that humans find easy but AI models often get wrong. Tests common sense and reasoning gaps.
Writing quality evaluation by Lech Mazur. Tests prose quality, coherence, and stylistic ability.
LiveBench fiction analysis. Tests literary comprehension and creative text understanding.
- Typetext
- Context164K tokens (~82 books)
- ReleasedAug 2025
- LicenseOpen Source
- StatusActive
- Cost / Message~$0.001
Frequently Asked Questions
Key facts · as of 2026-10-05
- DeepSeek V3.1 by DeepSeek. BenchGecko score 51.8, rank 131 of 312 scored models (normalized average of public benchmark scores).
- List price $0.25 input · $0.95 output per 1M tokens (as of 2026-10-05).
- Sold by 7 providers (as of 2026-10-05): DeepInfra (fp4) $0.25 in / $0.95 out · SiliconFlow (fp8) $0.27 in / $1.00 out · AtlasCloud (fp8) $0.30 in / $1.00 out · CoreWeave (fp8) $0.55 in / $1.65 out · Google $0.60 in / $1.70 out · and 2 more. Every provider
How to cite · data as of 2026-10-05
DeepSeek V3.1 · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/deepseek-chat-v3-1
Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP