DeepSeek
The open-weight family everyone else gets benchmarked against.
DeepSeek publishes the weights for nearly everything it ships, which is why its models became the price anchor for the whole market — a Flash model here costs a fraction of a frontier model and gets close on reasoning and code. The V4 line is a sparse mixture-of-experts design with a million-token context. This row is the model family; DeepSeek's own free chat app has its own entry in Chats.
At a glance — DeepSeek V4.1 Flash
- Cost
- $0.099 in · $0.60 out per 1M tokensBudget
- Memory
- 1M tokens — about 1,550 pages
- Reads
- Text · Images
- Can do
- Use tools (agents)
- Structured output
- Thinking mode
- Good at
- Overall: better than 75%
- Reliability
- 99.99% uptime, last 24h
- Released
- Sep 10, 2026
Advanced details
- API model ID
deepseek/deepseek-v4.1-flash- Max output
- 944K tokens
- Cached input
- $0.06 per 1M
- Tokenizer
- DeepSeek
- Overall index
- 39.5
Providers
The same model, hosted by different companies. OpenRouter picks one per request and falls back to the next if it fails.
| Provider | Input /1M | Output /1M | Cached /1M | Max output | Uptime 24h |
|---|---|---|---|---|---|
| InferenceNetDegraded | $0.035 | $0.29 | $0.001 | 384K | 58.36% |
| Relace | $0.09 | $0.45 | $0.009 | 944K | 99.88% |
| Wafer | $0.099 | $0.60 | $0.06 | 944K | 99.90% |
| OpenInference · fp4 | $0.10 | $0.50 | $0.01 | 944K | 99.55% |
| Morph | $0.12 | $0.48 | $0.002 | 944K | 99.77% |
| Sail Research · fp4 | $0.13 | $0.75 | $0.01 | 384K | 99.69% |
| DeepInfra · fp8 | $0.14 | $0.42 | $0.004 | 131K | 99.89% |
| DekaLLM | $0.15 | $0.60 | $0.005 | 944K | 98.89% |
| StreamLake · fp8 | $0.165 | $0.66 | $0.003 | 384K | 99.73% |
| CoreWeave · fp8 | $0.20 | $0.65 | $0.03 | 944K | 99.32% |
| NextBit · fp8 | $0.21 | $0.84 | $0.004 | 944K | 99.91% |
| Fireworks | $0.22 | $0.66 | $0.007 | 944K | 99.87% |
| Krea · fp8 | $0.225 | $0.90 | $0.006 | 944K | 97.34% |
| GMICloud · fp8 | $0.225 | $0.90 | $0.005 | 944K | 99.85% |
| Phala | $0.276 | $1.104 | $0.006 | 393K | 99.35% |
| Novita · fp8 | $0.285 | $1.14 | $0.006 | 393K | 99.99% |
| AtlasCloud · fp8 | $0.30 | $1.20 | $0.03 | 393K | 99.88% |
| BaseTen · fp8 | $0.30 | $1.20 | $0.03 | 33K | 99.48% |
| Makora · fp8 | $0.30 | $1.20 | $0.006 | 393K | 99.70% |
| DigitalOcean | $0.30 | $1.20 | $0.006 | 944K | 99.79% |
| Together | $0.30 | $1.20 | $0.006 | 944K | 99.90% |
| SiliconFlow · fp8 | $0.30 | $1.20 | $0.006 | 393K | 99.66% |
| Modal | $0.30 | $1.20 | $0.03 | 944K | 99.92% |
| BaseTen · fp8 | $0.30 | $1.20 | $0.007 | 33K | 98.68% |
| Parasail · fp8 | $0.30 | $1.20 | $0.006 | 944K | 98.93% |
| Venice · fp8 | $0.375 | $1.50 | $0.008 | 131K | 99.37% |
Supported parameters
- frequency_penalty
- include_reasoning
- logit_bias
- logprobs
- max_tokens
- min_p
- presence_penalty
- reasoning
- reasoning_effort
- repetition_penalty
- response_format
- seed
- stop
- structured_outputs
- temperature
- tool_choice
- tools
- top_k
- top_logprobs
- top_p
Live from OpenRouter, refreshed hourly. Scores from Artificial Analysis.
Key info
- Pricing
- Open source
- Category
- Models
- Best for
- Intermediate
- Made by
- DeepSeek
- Weights
- Open
- Good for
- Reasoning and code at a fraction of frontier prices