DeepSeek V3
deepseek/deepseek-chat- DeepSeek
- Family · DeepSeek V3
- Released 26 Dec 2024
- Open weights
- Text → Text
- Tools · JSON
Not on OSMEXO yet
Spec & list price
Context160K
Max output16K
Input / 1M$0.26
Output / 1M$1.034× input
Scores · standing
Intelligence8.5#294/489
Coding23.0#132/196
Usage
Tokens / day2.0B▼ 3.2% 7d
Usage rank#161
Requests / day430.6K
Tool calls / req0.02
Per request · 30 days
Avg input2.2K0.1× median
Avg output1740.3× median
Cache hit22%0.3× median
Tool-call failures3.9%2.0× median
Consumption
Peak 3.8B 9 OctLow 962.3M 31 OctRank today #161
News & trends
6 mentions this week ▼ 14%DeepSeek V3 family · videos, posts, articles · 30 days
Trending · 14 days · DeepSeek V3 family
- What Did DeepSeek V3's $5.6 Million Cost Leave Out? #deepseek #nvidia #ai #mixtureofexperts1.5K views
- Best provider for deepseek v3.2?17 upvotes
- Free Claude Code CLI Setup: Use DeepSeek V3 & Gemini API (1-Click Windows Installer)29 views
News · latest from DeepSeek
- Tell HN: Mistrap Pro includes DeepSeek v4.1 Flash 1017docs.mistral.ai · 1 Nov
- Show HN: DeepSeek v4.1 Flash-1017 Safety Evaluationtwitter.com · 26 Oct
- DeepSeek V4 Pro 0813 vs Claude Fable 5 on DeepSWE: Cost, Coding, and RoutingTogether AI · 17 Aug
- DeepSeek-V4 Flash 0731 vs GPT-5.6 Luna on DeepSWE: Cost and CodingTogether AI · 6 Aug
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 3 providers, latest day
Output speed#324/33314 tok/s4.8× slower than median
Fastest provider32 tok/sbest of 3 providers
First token#2331.65 s1.6× longer than median
#26 cheapest of 44 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases21.5
Technical details
- Tokenizer
- DeepSeek
- Knowledge cutoff
- 2024-07-31
- Parameters
- 684B
- Moderation
- none
- Downloads · likes
- 1.7M · 4.5K
Request parameters 15
- tools
- tool_choice
- response_format
- structured_outputs
- temperature
- top_p
- top_k
- min_p
- max_tokens
- stop
+5 more
- seed
- frequency_penalty
- logit_bias
- presence_penalty
- repetition_penalty
When to use
Best for
- Low-cost open-weight chat
- Math and general knowledge questions
- Tool-calling with JSON output
- Self-hosted general assistants
Think twice for
- Image input (text-only)
- Coding agents (coding scores are low)
- Long outputs (output cap is small)
- Hardest reasoning tasks
Research papers
- Beyond the Trial-and-Error Loop: Hybrid Projection and Automated Tuning for Distributed Training14 Oct
- Before You Poll with LLMs: A Deliberative Diagnostic Framework14 Sep
- FlowCheck: Helping End-Users Specify and Verify Intent in Vibe-Coded Web Apps28 Aug
- Revisiting the Performance of Generative Artificial Intelligence on Introductory Object-Oriented Programming Assessments: Insights from 202617 Aug
- DualityCert: Verifier-Gated Language-Model Repair of Broken Duality Claims in Quantum Field Theory26 Jul
- Decoding the Skew: Distribution-Aware MoE Inference with Adaptive Kernel Dispatch25 Jul
- Reading Between the Dots: Decoding Hidden Computation across Filler Tokens3 Jul
- Design and Evaluation of Multi-Agent AI Oracle Systems for Prediction Market Resolution29 May
Same family
DeepSeek V3- 13.7
- 21.5
- 16.6
- 9.7
Alternatives
other vendors · similar capability- 13.1
- 6.7
- 10.2
- 13.3
Want DeepSeek V3 on OSMEXO?
Sign in to request it. Requests are reviewed; a request does not guarantee availability.

