PUBLIC INTELLIGENCE
Selected AI market intelligence
Recent moves by company. Sign up for the full timeline, company profiles and trend analysis.
OpenAI Platform · major
GPT-5.6 integrated into OpenAI developer tool Kiro
GPT-5.6 is now available in OpenAI's developer tool Kiro, offering better cost-performance for software planning, building, and testing.
Volcano Engine Ark · major
Volcengine Ark Updates Quickstart Code with Responses API and New Model ID
Volcengine Ark's quickstart code examples are fully updated to use the date-suffixed model ID 'doubao-seed-2-1-pro-260628' and introduce the Responses API endpoint (/api/v3/responses). The code also includes a parameter to manually disable deep thinking (thinking=disabled), indicating the platform is adopting the new API paradigm.
Volcano Engine Ark · major
Volcano Engine Ark Adjusts DeepSeek V4 Series Model Pricing, Effective August 28
Volcano Engine Ark announced pricing adjustments for both DeepSeek V4 series models effective August 28, 2026. For deepseek-v4-flash, online inference input rises from 1 to 3 CNY/million tokens (+200%), output from 2 to 9 CNY (+350%), while cache hit drops from 0.2 to 0.1 CNY (-50%). For deepseek-v4-pro, input decreases from 12 to 9 CNY (-25%), output rises slightly from 24 to 27 CNY (+12.5%), and cache hit drops sharply from 1 to 0.3 CNY (-70%). The restructuring eliminates Flash as a low-cost entry point while incentivizing prompt optimization through aggressive cache pricing on Pro.
Gemini Enterprise Agent Platform · major
Gemini Enterprise Agent Platform Announces Component Billing and Unified Free Tier
Gemini Enterprise Agent Platform introduces a unified billing structure and component billing plans. Agent Compute, Memory, and Storage adopt unified billing with monthly free tiers, and idle runtime is not billed. Agent Gateway will be billed per vCPU-hour starting July 13, 2026; Memory Bank, Sessions, and Skill Registry will start storage and I/O billing on Sept 1, 2026. Gen AI Evals also updated pricing, defaulting to Gemini 2.0 Flash with lower computational metric rates.
OpenAI Platform · major
OpenAI Terminates Model Supply Contract with Cursor Following SpaceX Acquisition
OpenAI announced it will gradually terminate its model supply contract with Cursor following the AI coding tool's acquisition by SpaceX. As Cursor is a major player in AI coding, this move could impact its product roadmap and create opportunities for competitors like Anthropic and Google. OpenAI's decision may stem from competitive, data security, or strategic considerations to prevent core technology from flowing to potential rivals via acquisition. Existing Cursor users may face model switches or service changes, and enterprise clients relying on the Cursor+OpenAI stack must reassess their tech stack. Specific timelines and transition plans remain undisclosed.
Model Studio / Bailian · major
Alibaba Cloud Bailian Introduces Night Plan with 50% Off Qwen3.8-Max Inference
Alibaba Cloud Bailian launched the 'Night Plan' offering a 50% discount on Qwen3.8-Max inference during off-peak hours for specific subscription customers, optimizing compute utilization and reducing costs for developers.
Model Studio / Bailian · major
Alibaba Cloud adds personal tier and reduces team pricing for Token Plan
Alibaba Cloud's Qwen AI platform introduced a personal tier for its Token Plan and reduced pricing for the team tier. It also offers a 10x bonus and up to 80% off for Qwen3.8-Max during night hours, lowering the barrier for individual developers and intensifying price competition.
AWS Bedrock · major
AWS Bedrock Removes Grok 4.3 Pricing and Adds Grok 4.6 Entry
AWS Bedrock pricing page removed on-demand pricing details for Grok 4.3 in multiple US regions while adding a model selection entry for Grok 4.6. This suggests an adjustment in xAI's model lineup on Bedrock, with Grok 4.6 potentially replacing or supplementing Grok 4.3.
Gemini Enterprise Agent Platform · major
CodeMender Updates to Support Gemini 3.6/3.7 Flash and Adds Unrestricted Verification Flag
CodeMender released an update adding support for Gemini 3.6 Flash and 3.7 Flash (default) models. It introduces the --unrestricted flag for cm verify, allowing vulnerability exploitation verification bypassing command policy limits in isolated environments. It also fixes long-session reliability, cross-OS shell parsing, and permission denial error readability.
DeepSeek Platform · major
DeepSeek Implements Weekend Off-Peak Pricing and Clarifies Deduction Rules
DeepSeek adjusts its billing rules, implementing all-day off-peak pricing on weekends (Beijing time) starting August 23, 2026, to encourage large-scale batch inference and smooth out compute demand. It also clarifies deduction rules, prioritizing the use of gifted balances over recharged funds to alleviate user concerns about locked capital.
DeepSeek Platform · milestone
DeepSeek API Introduces Peak/Off-Peak Pricing with Significant Adjustments
DeepSeek announced a shift to peak/off-peak pricing for its API services, with off-peak rates at half the peak rates. Peak hours are UTC 01:00-04:00 and 06:00-10:00, effective August 16, 2026. For deepseek-v4-flash, off-peak input is $0.22/M and output $0.66/M; peak input is $0.44/M and output $1.32/M, representing a significant price increase during peak hours.
Zhipu BigModel · major
Zhipu BigModel Launches Independent Search Tool Service Line
Zhipu BigModel has introduced an independent 'Search Tool Service' product line with four tiers of search tools priced per query, along with a limited-time 40% discount. This decouples search capabilities from core models, forming a 'model + search plugin' combo and marking the first time a domestic model vendor independently prices search tools.
Zhipu BigModel · major
Zhipu BigModel Offers Limited-Time Free Access to All Flagship Models
Zhipu BigModel's pricing page now marks all flagship models as 'limited-time free', displaying both the promotional status and future standard pricing. This aggressive customer acquisition strategy aims to lower trial and migration costs, potentially shifting the domestic model API market's payment landscape in the short term.
Microsoft Azure AI Foundry · major
Azure Microsoft Foundry Launches Five New Agent Capabilities for Claude Models
Claude models hosted on Azure Microsoft Foundry gained five new capabilities: structured output, web search, web scraping, MCP connectors, and tool search, transforming model endpoints into production-grade agent platforms.
AWS Bedrock · major
AWS Bedrock Pricing Page Revamp and Removal of Multiple OpenAI Model Pricings
AWS Bedrock underwent a major structural revamp of its pricing page, replacing the detailed regional pricing tables for OpenAI frontier models with an interactive model selector. During this update, pricing for several older or specific OpenAI models was completely removed, including GPT-5.4 (across all regions including GovCloud), GPT-5.5, and Daybreak Red: GPT-5.6 Cyber. Furthermore, the previous regional inference pricing structure and cross-region inference preview notes were deleted. These changes indicate AWS is simplifying its pricing display logic, potentially accompanied by availability adjustments for OpenAI models on Bedrock or a migration push toward newer model generations.
Huawei Cloud ModelArts · major
ModelArts Notebook integrates multiple AI assistants and Skills
On August 13, 2026, ModelArts introduced best practices for integrating AI coding assistants like OpenClaw and Cline into Notebook.
The above is a public selection. Sign up to access full timelines, per-source evidence and trend analysis.
Sign up for more intelligence →