PUBLIC INTELLIGENCE
Selected AI market intelligence
A public selection of recent important updates. Create an account for the complete timeline, company profiles, and trend analysis.
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.
OpenAI Platform · major
OpenAI Previews Ultrafast Mode with 14x Speedup for GPT-5.6 Sol
OpenAI previewed the new Ultrafast API service tier, leveraging Cerebras chips to boost GPT-5.6 Sol's processing speed by up to 14x, reaching up to 750 tokens per second and significantly reducing inference latency.
DeepSeek Platform · major
DeepSeek V4 Series Model Upgrades, Responses API Support, and Concurrency Limit Adjustments
DeepSeek has rolled out significant updates to its V4 series, advancing DeepSeek-V4-Pro to version 0813 and natively supporting the Responses API to streamline complex Agent development. Concurrent request limits remain fixed at 2,500 for V4-Flash and 500 for V4-Pro, while legacy footnotes were removed to clarify usage. These changes elevate inference performance and expand the Agent development ecosystem.
DeepSeek Platform · major
DeepSeek Announces Imminent Comprehensive Price Increase for API Services
DeepSeek has announced plans to implement a significant, comprehensive price increase across its API services shortly, advising users to plan their consumption accordingly. As an industry pricing benchmark, this strategic shift reflects evolving compute costs and commercialization goals, which will likely recalibrate market expectations around API pricing structures. Developers and enterprises dependent on low-cost models will need to reassess the trade-offs between expenditure and capability, potentially sparking renewed competitive dynamics.
Vertex AI / Google Cloud · major
Google DeepMind Launches Gemini 3.7 Flash
Google DeepMind officially launched Gemini 3.7 Flash, the latest model in the Gemini Flash series, designed to deliver higher cost-effectiveness and faster inference capabilities.
DeepSeek Platform · major
DeepSeek-V4-Pro Officially Launched with Enhanced Agent Capabilities
DeepSeek-V4-Pro is officially GA on APP, Web, and API. This release significantly enhances agent capabilities, achieving outstanding results in multiple agent benchmarks, including Terminal Bench 2.1 (87.9), DeepSWE (62.7), and DSBench-FullStack (71.1).
DeepSeek Platform · major
DeepSeek API Natively Supports OpenAI Responses API Format
The DeepSeek API now natively supports the OpenAI Responses API format and is specifically adapted for Codex, providing one-click configuration scripts that significantly reduce the migration cost for developers moving from OpenAI to DeepSeek.
Huawei Pangu · major
ModelArts Studio Integrates DeepSeek R1/V3 and Offers Agent Development Capabilities
ModelArts Studio serves as a unified enterprise portal delivering inference services for Huawei's Pangu models alongside third-party offerings, including the full DeepSeek R1 and V3 series. The platform also provides integrated large model services spanning data engineering, model development, domain applications, and Agent creation.
Huawei Cloud ModelArts · major
Huawei Cloud ModelArts Inference Service Upgrades: Multi-Level Flow Control, Agentic RL, and Multi-P/Multi-D Deployment
ModelArts' inference service has been upgraded with a decoupled architecture supporting gray releases and traffic mirroring for smooth rollouts, alongside multi-level flow control and rapid recovery mechanisms. An integrated evolutionary reinforcement learning framework (Agentic RL) enables continuous agent evolution, while new multi-pod/multi-device deployment modes and topology visualization improve efficiency and operational visibility. These enhancements collectively strengthen SLA guarantees and developer experience for enterprise workloads.
Huawei Cloud ModelArts · major
ModelArts Supports Trillion-Parameter, Petabyte-Scale Distributed Training Within Single Jobs
ModelArts can execute ultra-large-scale distributed tasks, supporting training for trillion-parameter models and hundred-petabyte data volumes within a single job. The platform includes built-in operations capabilities such as performance profiling and automated fault diagnosis.
Huawei Cloud ModelArts · major
Huawei Cloud ModelArts Launches Industry-First One-Click Deployment of Ascend-Adapted DeepSeek-V4-Flash
Huawei Cloud ModelArts introduces the industry's first natively adapted DeepSeek-V4-Flash model for Ascend hardware, supporting 1 million token context inference with TPOT optimized to 20ms and significantly increased single-card throughput compared to DeepSeek-V3.2.
OpenAI Platform · major
AWS Bedrock Launches Daybreak Models and Updates GPT-5.6 Pricing and Regional Support
AWS Bedrock announced the availability of OpenAI’s Daybreak cybersecurity models and updated pricing and regional support across the GPT-5.6 series (including Sol, Terra, Luna, and Daybreak Red/Blue variants). The Ohio region added full GPT-5.6 support along with 5.5/5.4 pricing, while Northern Virginia introduced Daybreak Blue (GPT-5.6 Sol) and Daybreak Red (GPT-5.6 Cyber) pricing. Some workloads were migrated from Northern Virginia to Ohio to create a more granular, tiered pricing and regional matrix.