March 2024 iteration of the DeepSeek V3 685B MoE model.
xAI's new flagship for coding, agentic tasks, and knowledge work, succeeding Grok 4.6 with its longest reinforcement learning run yet.
No shared demos for these models yet.
No captured outputs for these models yet.
Both DeepSeek V3 (0324) and Grok 4.7 are capable AI models. The best choice depends on your specific use case: consider pricing, context window, speed, and which capabilities (vision, reasoning, audio) you need.
DeepSeek V3 (0324) has a lower blended cost. DeepSeek V3 (0324): $0.29 input / $1.14 output. Grok 4.7: $2.00 input / $6.00 output.
Grok 4.7 has a larger context window: DeepSeek V3 (0324) supports 128K tokens vs Grok 4.7 at 500K tokens.
Grok 4.7 supports vision/image input, but DeepSeek V3 (0324) does not.
Key differences: DeepSeek V3 (0324) is significantly cheaper; only Grok 4.7 supports vision input; only Grok 4.7 offers extended reasoning mode. Compare full specs on this page.
If cost is your priority, choose the cheaper option. If you need the highest intelligence for complex tasks, pick the higher-scoring model. For long documents or codebases, choose the larger context window. You can try both DeepSeek V3 (0324) and Grok 4.7 for free on idapt.app to see which performs better for your specific needs.