Question 1

What is DeepSeek V3?

Accepted Answer

DeepSeek V3 is a model from Deepseek, released on Dec 26, 2024. DeepSeek-V3 is a 671B-parameter Mixture-of-Experts language model from DeepSeek with 37B activated parameters per token, trained on 14.8T tokens.

Question 2

How much does DeepSeek V3 cost?

Accepted Answer

On the Respan gateway, input starts at $0.14 per 1M tokens on Azure DeepSeek, and output starts at $0.28 per 1M tokens on Azure DeepSeek, after gateway discounts. Averaged across 7 providers, it costs $0.33 per 1M input tokens and $0.86 per 1M output tokens.

Question 3

Which providers serve DeepSeek V3 on Respan?

Accepted Answer

7 providers serve DeepSeek V3: Azure DeepSeek, OpenRouter, DeepInfra, Featherless, and Novita. Azure DeepSeek require your own provider API key (BYOK).

Question 4

What is DeepSeek V3's context window?

Accepted Answer

DeepSeek V3 accepts up to 164K tokens of context on its largest provider route.

Question 5

How do I use DeepSeek V3 with the Respan gateway?

Accepted Answer

Point any OpenAI-compatible SDK at https://api.respan.ai/api/ with your Respan API key and set the model to "deepseek-v3". To set it up with a coding agent, copy the agent setup prompt at the top of this page.

Question 6

How fast is DeepSeek V3?

Accepted Answer

Over the last 30 days of requests through the Respan gateway, the fastest provider was Novita at 14 tokens per second, and the lowest response latency was 0.9 seconds on Featherless. Speed varies with request size and provider.

Question 7

Where does DeepSeek V3 rank on the Respan Index?

Accepted Answer

DeepSeek V3 scores 12.4 out of 100 on the Respan Index v1.1, ranking #132 of 138 models. By category: Coding 3.7, Agents 7.9 (estimated), Reasoning 9.8, Knowledge & language 5.9, and Human preference 43.2. The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by llmmetric.com.

Question 8

How does DeepSeek V3 score on benchmarks?

Accepted Answer

DeepSeek V3 has 11 published benchmark scores, including SWE-bench Verified 42%, GPQA Diamond 67.6%, and MATH Level 5 75.5%. Each score links to where it was published; the data is collected by llmmetric.com.

Provider	Context	Input	Output
Azure DeepSeekBYOK only`azure_deepseek/deepseek-chat`	64K	$0.14/M	$0.28/M
Azure DeepSeekBYOK only`azure_deepseek/deepseek-v3`	64K	$0.14/M	$0.28/M
OpenRouter`openrouter/deepseek/deepseek-chat`	164K	$0.26/M	$1.03/M
OpenRouter`openrouter/deepseek/deepseek-chat-v3-0324`	164K	$0.29/M	$1.14/M
DeepInfra`deepinfra/deepseek-ai/DeepSeek-V3`	164K	$0.32/M	$0.89/M
Featherless`featherless_ai/deepseek-ai/DeepSeek-V3-0324`	131K	$0.39/M	$1.63/M
NovitaSave 10%`novita/deepseek/deepseek_v3`	64K	$0.80/M	$0.80/M

DeepSeek V3

Providers

Certified where it counts.

Pricing

Cost by volume

Throughput and latency

Output speed

Response latency

Benchmarks

Respan Indexv1.1

Coding2

Reasoning

Apps

DeepSeek V3 FAQ

Works with your entire stack

Built for AI agents.
Break less.
Ship more.

Math3

Knowledge2

Arena1