DeepSeek: DeepSeek V3
deepseek/deepseek-chatDeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.
Modalities
In / out per 1M$0.32 / $0.89
Context160K
ReleasedDec 26, 2024
Providers
2 providers serve this model upstream.
| Provider | Input /M | Output /M | Cache read /M | Latency P50 | Throughput P50 | Uptime |
|---|---|---|---|---|---|---|
| $0.26 | $1.03 | Variable | 1.46s | 29 tps | 99.33% | |
 DeepInfraus · fp4 | $0.32 | $0.89 | Variable | 0.57s | 18 tps | 88.86% |
Market-wide measurements, refreshed every 15 minutes. They describe who serves DeepSeek V3 in general, not how ProjectCOZY routes your requests.
Pricing
List rates per 1M tokens — what DeepSeek V3 costs pay-as-you-go elsewhere, not what you pay here.
Input
$0.32
per 1M tokens
Output
$0.89
per 1M tokens
Cache read
—
not published
Included in every pass
DeepSeek V3 is included in every ProjectCOZY pass at no extra cost — one flat rate, no per-token metering. The rates above are there for comparison.
Performance
Time to first token and output speed across every provider, over the last 3 days.
0.0s2.0s4.0s6.0s8.0s
Oct 11, 12 AMOct 11, 8 AMOct 11, 4 PMOct 12, 12 AMOct 12, 8 AMOct 12, 4 PMOct 13, 12 AMOct 13, 8 AMOct 13, 4 PMOct 14, 12 AMOct 14, 8 AM
DeepInfra (us) StreamLake
0 tok/s20 tok/s40 tok/s60 tok/s80 tok/s
Oct 11, 12 AMOct 11, 8 AMOct 11, 4 PMOct 12, 12 AMOct 12, 8 AMOct 12, 4 PMOct 13, 12 AMOct 13, 8 AMOct 13, 4 PMOct 14, 12 AMOct 14, 8 AM
DeepInfra (us) StreamLake
Uptime
100% of requests were served successfully over the last 3 days, across all providers.
Last 3 daysAvailability 99.93%
Oct 11, 2:00 PMNow
Availability over the last 24 hoursAvailability 99.92%
100% 98% 97% 95%
Oct 13, 2:40 PMNow
Apps
Tools that most often run DeepSeek V3, by tokens this week.
Activity
88.0B tokens processed market-wide in the last month.
01.0B2.0B3.0B4.0B
Sep 14Sep 19Sep 24Sep 29Oct 4Oct 9Oct 14
Tokens
FAQ
Which model id do I use for DeepSeek V3?
deepseek/deepseek-chat. Send it as the model field from any OpenAI-compatible client, against http://192.168.1.55/v1.
How much context does DeepSeek V3 hold?
163,840 tokens per request, of which up to 16,384 can be the response.
What can it accept and return?
Input: text. Output: text.
What does DeepSeek V3 cost on a pass?
Nothing beyond the pass itself. Pay-as-you-go elsewhere it lists at $0.32 per 1M input tokens and $0.89 per 1M output tokens.
How current is its knowledge?
Training data runs to Jul 2024. Anything later has to come from your prompt or a tool call.
Who built DeepSeek V3?
DeepSeek, in the DeepSeek family.
Explore
Other models your pass can call.
