This model is deprecated. We still benchmark its performance, but less often: every 18 hours instead of every 6.
DeepSeek has launched a newer release, DeepSeek R1 0528 (May '25). We suggest considering it instead.
DeepSeek R1 (Jan '25): Intelligence, Performance & Price Analysis
Available providers
Model summary
Verbosity
DeepSeek R1 (Jan '25) supports text input, outputs text, and has a 128k tokens context window.
Pricing for DeepSeek R1 (Jan '25) is $2.00 per 1M input tokens (expensive, median: $0.43) and $4.00 per 1M output tokens (expensive, median: $1.84).
| Reasoning | Yes |
|---|---|
| Input modality | Supports: text |
| Output modality | Supports: text |
| Context window | 128k ~192 A4 pages of size 12 Arial font |
| Total parameters | 685B |
| Active parameters | 37B Number of parameters active per token during inference |
| License | MIT |
| Model weights | Hugging Face |
Metrics are compared against models of the same class:
- Non-reasoning models → compared only with other non-reasoning models
- Reasoning models → compared across both reasoning and non-reasoning
- Open weights models → compared only with other open weights models of the same size class:
- Tiny: ≤4B parameters
- Small: 4B–40B parameters
- Medium: 40B–150B parameters
- Large: >150B parameters
- Proprietary models → compared across proprietary and open weights models of the same price range, using a blended 3:1 input/output price ratio:
- <$0.15 per 1M tokens
- $0.15–$1 per 1M tokens
- >$1 per 1M tokens
Intelligence
Artificial Analysis Intelligence Index
Artificial Analysis Intelligence Index by Open Weights / Proprietary
Intelligence Index vs. Cost per Task
Evaluations
Agentic knowledge work, Elo
Agentic SaaS workflows
Professional document reasoning, All-pass
Agentic coding & terminal use
Agentic real-world work tasks, Elo
Openness Index
Artificial Analysis Openness Index: Score
Cost
Cost per Intelligence Index Task
Cost to Run Artificial Analysis Intelligence Index
Pricing: Cache Hit, Input, and Output
Performance
Measured by time per task (minutes), output speed (tokens per second), time to first answer token and time to first token (seconds)
Time per Intelligence Index Task
Output Speed
Time to First Answer Token
Context Window
Context Window
Model Size (Open Weights Models Only)
Model Size: Total and Active Parameters
Further details
Weights | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek R1 (Jan '25) | - | 685B 37B active at inference time | 128k | $2.20 | - | 129 |
Frequently Asked Questions
Common questions about DeepSeek R1 (Jan '25)
DeepSeek R1 (Jan '25) was released on January 20, 2025.
DeepSeek R1 (Jan '25) was created by DeepSeek.
DeepSeek R1 (Jan '25) has a time to first answer token of 13.02s (based on the median across providers serving the model), which is better than average compared to other open weight models of similar size (median: 19.92s).
DeepSeek R1 (Jan '25) costs $2.00 per 1M input tokens (at the higher end, median: $0.43) and $4.00 per 1M output tokens (at the higher end, median: $1.84), based on the median across providers serving the model.
DeepSeek R1 (Jan '25) costs $2.00 per 1M input tokens and $4.00 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 $2.20 per 1M tokens. Pricing may vary by provider. Compare provider pricing
Yes, DeepSeek R1 (Jan '25) is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.
DeepSeek R1 (Jan '25) supports text input.
DeepSeek R1 (Jan '25) supports text output.
No, DeepSeek R1 (Jan '25) does not support image input. It can only process text.
No, DeepSeek R1 (Jan '25) is not multimodal. It only supports text input.
DeepSeek R1 (Jan '25) has a context window of 128k tokens. This determines how much text and conversation history the model can process in a single request.
Yes, DeepSeek R1 (Jan '25) is open weights. The model weights are publicly available and can be downloaded for self-hosting.
DeepSeek R1 (Jan '25) has 685B parameters (37B active).
DeepSeek R1 (Jan '25) is a Mixture of Experts (MoE) model with 685B total parameters, but only 37B active parameters are used during inference.
DeepSeek R1 (Jan '25) is released under the MIT license. This license allows commercial use. View license
Yes, DeepSeek R1 (Jan '25) is available via API through 4 providers. Compare API providers
DeepSeek R1 (Jan '25) is available through 4 API providers. Compare providers