Deepseek Flash Challenges GPT-5.6 Luna with 60% Lower Cost

Deepseek Flash Challenges GPT-5.6 Luna with 60% Lower Cost

The new model Deepseek V4 Flash "0731" competes with OpenAI's GPT-5.6 Luna at a cost approximately 60% lower.

Deepseek has launched a major update to its cost-effective AI model. According to the Artificial Analysis Intelligence Index, the new version scores 50 points, ten points higher than the previous V4 Flash released in April 2026. This places it just one point behind OpenAI's economic model, GPT-5.6 Luna, but its cost per task is about 60% less, even after OpenAI's 80% price reduction. One of the main reasons for this gap is the 98% discount on Deepseek's cache, well above the industry standard of 90%. Additionally, the model uses 12% fewer tokens than its predecessor.

The Artificial Analysis Intelligence Index shows that the Deepseek V4 Flash "0731" achieves a score of 50 points after its update, coming very close to OpenAI's GPT-5.6 Luna while claiming the top spot for value for money.

The model improves across all tested categories compared to the previous version, with the most significant gains in agentic tasks. On GDPval, a benchmark designed to test models on complex office tasks, it rises from 1,189 to 1,559 Elo points. It also hallucinates less frequently. The architecture remains the same: 284 billion parameters in total, 13 billion active, with a context window of one million tokens. The model weights are available under an MIT license on Hugging Face.