Updated pricing data from pricepertoken.com shows DeepSeek V4 Flash has dropped to $0.09 per million input tokens and $0.18 per million output tokens as of its April 2026 release, even lower than the $0.14/$0.28 figures cited previously. Against OpenAI's GPT-5.5 at $5/$30, this represents a roughly 56x difference on input tokens and 167x difference on output tokens — gaps that have widened rather than narrowed since the last check.
A June 25, 2026 Axios report adds another data point: Z.ai's GLM-5.2 offers agentic capabilities rivaling Claude Opus 4.8 and GPT-5.5 while costing roughly half as much to run. This is significant because it shows the pricing advantage extends beyond budget-tier Chinese models to frontier-capable offerings. The report notes that GLM-5.2's open-weight nature adds additional appeal for developers who want to self-host and fine-tune.
A TechRepublic article from June 26, 2026 confirms the broader trend, reporting that Chinese AI models continue to undercut OpenAI and Anthropic on price, making them increasingly hard for US developers and startups to ignore. The article frames the cost advantage as particularly relevant for teams running large volumes of prompts, code tasks, summaries, or internal search queries.
The price advantage stems from deliberate Chinese market strategies. As documented in the ongoing AI price war, providers like Xiaomi have slashed API costs by up to 99%, and DeepSeek has made its discount campaigns permanent. These moves target enterprise developers processing billions of tokens daily, where per-token costs become the critical metric.
The data continues to justify the "unbeatable" and "different league" characterization for the price floor. Even the cheapest Western model, Google's Gemini 3.1 Flash-Lite at $0.25/$1.50, costs roughly 2.8x more on input and 8.3x more on output than DeepSeek V4 Flash. The new data strengthens rather than weakens the original verdict.
One caveat noted by TechRepublic: while token prices are dramatically lower, enterprise adoption depends on more than cost — including data governance, security review, compliance, support, and vendor risk. These factors may slow uptake among risk-averse organizations regardless of pricing.
