Back to Devoured - August 13, 2026
DEVOURED
deepseek prices its new v4 pro 0813 model at 0 87 per 1 million output tokens as the high flying chinese ai lab wows with its soaring token consumption

deepseek prices its new v4 pro 0813 model at 0 87 per 1 million output tokens as the high flying chinese ai lab wows with its soaring token consumption

AI llmfintech Wccftech
DeepSeek released the V4-Pro-0813 model, pricing it at $0.87 per million output tokens to remain competitive in the ongoing AI price war.
What: The V4-Pro-0813, which handles 1 million context tokens, is positioned against OpenAI's GPT-5.6 Luna and Anthropic's Opus 4.8, reportedly outperforming them on benchmarks like Terminal Bench 2.1.
Why it matters: The aggressive pricing and high volume of token consumption suggest DeepSeek is leveraging its position as a high-volume provider to force competitors into sustained price reductions.
Deep dive
  • DeepSeek's V4-Pro model utilizes 1.6 trillion parameters with 49 billion active parameters.
  • It currently holds strong performance on terminal-based benchmarks and automation tasks.
  • The model faces infrastructure pressure, with reports of inference speeds slowing due to high demand on a limited supply of approximately 20,000 NVIDIA H100 GPUs.
  • Pricing is set at $0.435/1M input and $0.87/1M output tokens.
Decoder
  • Inference: The process where a trained AI model processes input data to generate predictions or content.
  • Parameters: The internal variables or 'weights' of a neural network that determine its behavior and 'intelligence'; generally, more parameters allow for higher complexity.
Original article

DeepSeek Prices Its New V4-Pro-0813 Model At $0.87 Per 1 Million Output Tokens, As The Chinese AI Lab Comes Out Second Only To Anthropic On Token Consumption

DeepSeek was second only to Anthropic in terms of the total number of tokens consumed in July. And now, perhaps in a bid to cement its ascendancy, the high-flying Chinese AI lab has just unveiled the DeepSeek-V4-Pro-0813 model, its latest gambit to take on the might of OpenAI and Anthropic.

The V4-Pro-0813 is now rolling out on DeepSeek API and DeepSeek Chat

As yet unverified DeepSeek V4 Pro benchmarks from WeChat. Seismic if accurate. https://t.co/fEOONbaFEO pic.twitter.com/CG5s8Q1PJJ

— Andrew Curran (@AndrewCurran_) August 12, 2026

DeepSeek has started rolling out the V4-Pro on its API and Chat. The AI lab has priced the model at $0.435 per 1 million tokens of input and $0.87 per 1 million tokens of output.

Do note that OpenAI launched a literal price war a few days back by discounting its GPT-5.6 Luna by as much as 80 percent, with input tokens now priced at just $0.2 per 1 million from their earlier perch at $1, and output tokens priced at just $1.20 per 1 million vs. the earlier price of $6.

Just hours later, however, DeepSeek launched a refreshed version of its latest Flash-class model, dubbed the V4-Flash-0731. Critically, the model has just 284 billion parameters and yet offers a performance that is similar to Anthropic's Opus 4.8, which is widely believed to span multi-trillion parameters!

And, in what went right to the heart of OpenAI's price war, DeepSeek priced the V4 Flash 0731 at just $0.14 per 1 million tokens of input, and $0.28 per 1 million tokens of output, eviscerating any comparative price advantage that OpenAI tried to garner with its discounting move.

DeepSeek silently released V4-Pro 0813, up 15.8% on Terminal Bench from their April Preview model, with Fable 5 performance at ~57x cheaper cost.

1.6T param, 49B active, 1M context. This is the best price-to-perfomance model on the market right now.

Available in ClinePass now! pic.twitter.com/D9yas0umPn

— Cline (@cline) August 12, 2026

Coming back, DeepSeek's V4-Pro model outcompetes Opus 4.8 on Terminal Bench 2.1, Cybergym, DeepSWE, and AutomationBench benchmarks, as per the preliminary results populating WeChat right now.

Amazing stuff here. DeepSeek was 2nd only to Antropic in token volume in July. This was b4 V4-Flash 0731 came out. I'd imagine it will be 1st for August.

Sure, it's behind both Moonshot & Zai in revenue, but that's a lot of data DS is getting for post training. If DS can keep px… https://t.co/QKyGBiYK38 pic.twitter.com/Ql9wnHq2nn

— tphuang (@tphuang) August 12, 2026

Meanwhile, as stated earlier, DeepSeek was second only to Anthropic in terms of token volume in July, and might even clinch the apex spot in the coming months.

As such, DeepSeek is currently contending with an unprecedented demand surge, especially amid anecdotes that suggest its models' inference speeds slow down to a crawl at times, which is wholly understandable given the lab's limited compute footprint of just around 20,000 NVIDIA H100 GPUs.

Explore all 54 articles in today's digest

Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi