Model Library
DeepSeekReleased December 26, 2024DeepSeek V3 (Dec)
DeepSeek · Released December 26, 2024
Intelligence score
14
At a glance
- Intelligence
- 14
- Speed
- —
- Latency
- —
- Input price
- $0.40per 1M tokens
- Output price
- $0.89per 1M tokens
- Context window
- 130Ktokens
- Parameters
- 671billions
Summary
Comparison summary
DeepSeek V3 (Dec '24) scores 14 on the Artificial Analysis Intelligence Index, placing it below average among other open weight non-reasoning models of similar size (median: 17). DeepSeek V3 (Dec '24) costs $0.40 per 1M input tokens (better than average, median: $0.60) and $0.89 per 1M output tokens (very competitive, median: $2.40), based on the median across providers serving the model.
Specifications
Technical specifications
| Model name | DeepSeek V3 (Dec) |
|---|---|
| Provider | DeepSeek |
| Release date | December 26, 2024 |
| Intelligence score | 14 |
| Speed | — |
| Latency (TTFT) | — |
| Input price | $0.40 / 1M tokens |
| Output price | $0.89 / 1M tokens |
| Context window | 130,000 tokens |
| Parameters | 671B |
| Reasoning | No |
| Input modalities | text |
| Output modalities | — |
| Open weights | Yes |
| License | DEEPSEEK LICENSE AGREEMENT |
| API availability | — |
| Source | Model Intelligence legacy profile: deepseek-v3 |
Pricing
Token prices
Price per 1M tokens
DeepSeek V3 (Dec), as listed in the catalog record
- Input$0.40
- Output$0.89
| Token type | Price per 1M |
|---|---|
| Input | $0.40 |
| Output | $0.89 |
Model Intelligence legacy profile: deepseek-v3
FAQ · 19
Frequently asked questions
DeepSeek V3 (Dec '24) was released on December 26, 2024.
DeepSeek V3 (Dec '24) was created by DeepSeek.
DeepSeek V3 (Dec '24) scores 14 on the Artificial Analysis Intelligence Index, placing it below average among other open weight non-reasoning models of similar size (median: 17).
DeepSeek V3 (Dec '24) costs $0.40 per 1M input tokens (better than average, median: $0.60) and $0.89 per 1M output tokens (very competitive, median: $2.40), based on the median across providers serving the model.
DeepSeek V3 (Dec '24) costs $0.40 per 1M input tokens and $0.89 per 1M output tokens (based on the median across providers serving the model). For a blended rate (7:2:1 cache hit/input/output ratio), this is $0.45 per 1M tokens. Pricing may vary by provider.
When evaluated on the Intelligence Index, DeepSeek V3 (Dec '24) generated 3.3M output tokens, which is very competitive compared to other open weight non-reasoning models of similar size (median: 4.8M).
No, DeepSeek V3 (Dec '24) is not a reasoning model. It provides direct responses without extended chain-of-thought reasoning.
DeepSeek V3 (Dec '24) supports text input.
DeepSeek V3 (Dec '24) supports text only output.
No, DeepSeek V3 (Dec '24) does not support image input. It can only process text.
No, DeepSeek V3 (Dec '24) is not multimodal. It only supports text input.
DeepSeek V3 (Dec '24) has a context window of 130k tokens. This determines how much text and conversation history the model can process in a single request.
Yes, DeepSeek V3 (Dec '24) is open weights. The model weights are publicly available and can be downloaded for self-hosting.
DeepSeek V3 (Dec '24) has 671 billion parameters (37 billion active).
DeepSeek V3 (Dec '24) is a Mixture of Experts (MoE) model with 671 billion total parameters, but only 37 billion active parameters are used during inference.
DeepSeek V3 (Dec '24) is released under the DEEPSEEK LICENSE AGREEMENT license. This license allows commercial use.
DeepSeek V3 (Dec '24) achieves a score of 14 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.
Yes, DeepSeek V3 (Dec '24) is available via API through 5 providers.
DeepSeek V3 (Dec '24) is available through 5 API providers.