Alibaba, DeepSeek Drive China’s AI Model Race to Lower Costs

Alibaba has introduced Qwen3.8-Max, its most powerful AI model yet, as DeepSeek’s new V4-Flash gains attention for offering inference pricing lower than several competitors.

With 2.4 trillion parameters, Qwen3.8-Max employs a mixture-of-experts architecture that activates only a subset of parameters per request. Alibaba reports that approximately 95 billion parameters are active at any given time, significantly cutting costs and latency compared to models that activate all parameters.

DeepSeek utilizes a similar sparse architecture but on a smaller scale. According to Artificial Analysis, V4-Flash contains 284 billion total parameters with 13 billion active during inference, while Moonshot AI’s Kimi K3 features 2.8 trillion total parameters and roughly 104 billion active ones.

Qwen3.8-Max handles text, images, and video, supporting a context window of up to one million tokens. Alibaba also noted that the model successfully completed a software engineering project spanning 16 days.

In terms of size, Qwen3.8-Max rivals Kimi K3, released by Moonshot AI in July. The two firms are also competing on pricing: Qwen3.8-Max charges $2 per million input tokens and $6 per million output tokens, whereas Kimi K3 costs $3 and $15, respectively.

Model size alone does not dictate inference costs. Factors such as architecture, active parameter count, token consumption, and the number of calls needed to complete a task all influence the final cost of running a model.

Following its release, Qwen3.8-Max climbed to the top spot among Chinese text models on crowdsourced platform Arena.AI, though it still trails several Anthropic models in overall rankings. It also secured second place on Arena.AI’s leaderboard for image and visual analysis, behind an Anthropic Claude Fable 5 variant.

DeepSeek pushes down inference pricing

DeepSeek has adopted a different strategy with V4-Flash. Instead of matching the massive scale of Alibaba’s and Moonshot AI’s latest models, it has priced the model below several widely used AI systems.

According to Artificial Analysis, V4-Flash costs $0.14 per million input tokens and $0.28 per million output tokens. The research firm lists the model with a one-million-token context window and 284 billion total parameters, of which 13 billion are active during inference.

Artificial Analysis lists cache-hit pricing for the Max Effort version of V4-Flash at $0.003 per million tokens, which is 98% lower than its standard input rate. Cached input refers to previously processed context that can be reused across subsequent requests.

DeepSeek’s lower token rates also translated to Artificial Analysis’ benchmark testing. Reuters reported that the research firm estimated V4-Flash’s average cost at three cents per test, compared with 86 cents for Kimi K3, $1.86 for OpenAI’s GPT-5.6 Sol, and $3.15 for Anthropic’s Claude Fable 5.

This comparison accounts for the amount of input and output each model uses to complete the benchmark. A lower per-token rate does not necessarily mean a lower task cost if a model generates more output or requires additional interactions.

Artificial Analysis gave the Max Effort reasoning version of DeepSeek V4-Flash a score of 40 on its Intelligence Index. The research firm also recorded an output rate of approximately 118 tokens per second during testing.

Token prices tell only part of the cost story

Moonshot AI’s Kimi K3 offers another example of how advertised API prices can differ from the cost of completing longer workloads. Artificial Analysis lists the model at $3 per million input tokens and $15 per million output tokens, with cached input priced at $0.30 per million tokens.

On Artificial Analysis’ AA-Briefcase benchmark for agentic knowledge work, Kimi K3 averaged $10.57 per task. It generated around 120,000 output tokens and used an average of 83 turns per task.

Artificial Analysis stated that the cost reflected Kimi K3’s token pricing, output volume, and number of model interactions. Repeated model calls and larger outputs can therefore raise the total cost of completing a workload beyond what the headline API rate suggests.

Kimi K3 recorded the second-highest overall score on the AA-Briefcase evaluation at the time of testing, behind Claude Fable 5. It also scored 57 on Artificial Analysis’ broader Intelligence Index.

The comparison with DeepSeek highlights why cost-per-task measurements add useful context to standard API pricing. Models with different architectures and usage patterns can consume substantially different amounts of compute and tokens while working through the same type of task.

Open weights add another deployment option

Cost is also being shaped by how Chinese developers distribute their models. Alibaba, DeepSeek, and Moonshot AI have continued to support open-weight releases alongside hosted API access, giving developers more options for deploying these models.

Artificial Analysis lists DeepSeek V4-Flash as an open-weight model under an MIT licence, with weights available through Hugging Face. Kimi K3 is also available as an open-weight model under Moonshot AI’s own licence.

Open weights allow developers to run models on their own infrastructure or through third-party providers instead of relying solely on a developer-hosted inference service. Deployment costs still depend on the hardware and infrastructure used, but access to the model is not tied to a single hosted API.

This approach differs from the main models offered by OpenAI, Anthropic, and Google, which generally keep their model weights closed.

Lian Jye Su, chief analyst at Omdia, noted that model selection for many business workloads does not depend solely on having access to the highest-performing system.

“Many business workflows do not need the industry’s very best model,” Su said. “They need models that are good enough, affordable, transparent and accessible, and open-weight models help meet that demand.”

See also: Alibaba is designing AI chips around agents, and that changes what the race is actually about

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.