AI Model Pricing Shifts: NVIDIA, MoonshotAI, DeepSeek Cut Costs; Z.ai Adds Long-Context Model

Several AI model providers have announced pricing adjustments and new model releases. NVIDIA's Nemotron 3 Ultra has seen a completion price drop, benefiting long-form generation workloads. MoonshotAI's Kimi K2.7 Code and DeepSeek's V4 Flash models have also reduced prompt and completion costs, targeting developers sensitive to input token expenses and users seeking low-latency inference. Additionally, Z.ai has introduced GLM 5.2, a model with a 1,048,576 token context window, albeit with moderate-to-high generation costs. AI

IMPACT Price reductions and new models with varying context lengths offer more options for developers optimizing cost and performance.

RANK_REASON This item is a digest of pricing changes and new model additions from various AI providers, not a primary release from a frontier lab.

Read on dev.to — LLM tag →

• DeepSeek
• DeepSeek V4 Flash
• GLM 5.2
• Granite 4.0 Micro
• IBM
• inclusionAI
• Kimi K2.7 Code
• Ling-2.6-flash
• Llama 3.1 8B Instruct
• Meta
• MoonshotAI
• Nemotron 3 Ultra
• NVIDIA
• Z.ai

• product
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →