Skip to main content
​
2026.03.26

​
Product Updates

Account & BillingNew Features
  • Bonus Balance Migrated to Coupons All unused Bonus Balance has been automatically converted into platform-wide coupons. This is a format change only — your balance remains unchanged, and no action is required.
  • Unified Coupon Usage Coupons can now be used across all models and services on the platform, without any usage restrictions.
  • Coupon Expiration Introduced Coupons now have an expiration date. You can view the validity period in the Account Overview.
  • Overdue Account Restriction If your account balance is negative and no credit line is configured, your account will be considered overdue. In this state, model services and coupons will be temporarily unavailable. You can restore access by topping up your balance to a non-negative amount.
Improvements
  • Enhanced Usage & Billing Details You can now access more granular usage data and cost breakdowns.
  • Invoice Export Billing data can now be exported for easier reconciliation and financial management.
​
2026.03.11

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on March 17, 2026:
  • MiniMaxAI/MiniMax-M2.1
  • moonshotai/Kimi-K2-Thinking
  • Qwen/Qwen3-Next-80B-A3B-Thinking
  • Qwen/Qwen3-Next-80B-A3B-Instruct
  • meta-llama/Meta-Llama-3.1-8B-Instruct
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2026.03.05

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on March 12, 2026:
  • tencent/Hunyuan-MT-7B
  • black-forest-labs/FLUX.1 Kontext
  • Qwen/Qwen2.5-32B-Instruct
  • THUDM/GLM-4-9B-0414
  • THUDM/GLM-Z1-32B-0414
  • Qwen/Qwen3-Omni-30B-A3B-Thinking
  • deepseek-ai/DeepSeek-R1-Distill-Qwen-14B
  • Qwen/Qwen3-Omni-30B-A3B-Captioner
  • THUDM/GLM-Z1-9B-0414
  • Qwen/Qwen3-Omni-30B-A3B-Instruct
  • Qwen/Qwen3-VL-8B-Thinking
  • deepseek-ai/deepseek-vl2
  • Qwen/Qwen2.5-VL-7B-Instruct
  • Qwen/Qwen3-Reranker-4B
  • zai-org/GLM-4.5V
  • THUDM/GLM-4-32B-0414
  • Qwen/Qwen2.5-14B-Instruct
  • inclusionAI/Ling-mini-2.0
  • Qwen/Qwen2.5-Coder-32B-Instruct
  • Qwen/Qwen2.5-72B-Instruct-128K
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2026.02.08

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on Feb. 08, 2026:
  • MiniMaxAI/MiniMax-M2
  • MiniMaxAI/MiniMax-M1-80k
  • moonshotai/Kimi-Dev-72B
  • THUDM/GLM-4.1V-9B-Thinking
  • deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
  • Qwen/Qwen3-30B-A3B
  • stepfun-ai/step3
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2025.12.31

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on Dec. 31, 2025:
  • zai-org/GLM-4.5
  • Qwen/Qwen3-235B-A22B
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2025.11.19

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on Nov. 20, 2025:
  • inclusionAI/Ring-1T
  • inclusionAI/Ling-1T
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2025.09.25

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on Sept. 30, 2025:
  • Wan-AI/Wan2.1-I2V-14B-720P
  • Wan-AI/Wan2.1-I2V-14B-720P-Turbo
  • Wan-AI/Wan2.1-T2V-14B
  • Wan-AI/Wan2.1-T2V-14B-Turbo
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2025.09.04

​
Scheduled System Upgrade

​
Overview:

To deliver improved service quality, SiliconFlow will perform a scheduled system upgrade.
  • Date: Tuesday, September 9, 2025
  • Time: 02:00–10:00 UTC

​
Impact:

  • Login and account operations will be temporarily unavailable.
  • Model API calls will remain unaffected and can continue to be used normally.

​
Recommendations for Users:

  • Please ensure your account balance is sufficient prior to the maintenance window to avoid any service disruption.
  • Plan any critical login or account operations outside of the maintenance window.

​
Contact:

For any questions or support during the maintenance period, please contact contact@siliconflow.com.
​
2025.07.21

​
Platform Service Update

To better serve developers around the world, SiliconFlow has officially launched a new API endpoint: api.siliconflow.com. The existing endpoint api.siliconflow.com will remain available as well.If you have any questions, feel free to contact us.

​
Model Deprecation Notice

To further optimize resource allocation and deliver more advanced and high-quality services, the following models will be deprecated on July 28, 2025:
  • meta-llama/Llama-3.3-70B-Instruct
  • deepseek-ai/DeepSeek-R1-Distill-Llama-70B
  • deepseek-ai/DeepSeek-R1-Distill-Llama-8B
If you are currently using any of these models, we recommend switching to alternative models as soon as possible to avoid any service disruptions.
​
2025.05.30

​
Platform Service Adjustment Notice

As of now, the deepseek-ai/DeepSeek-R1 model has been updated to the latest version, 0528. You can still access the previous version via deepseek-ai/DeepSeek-R1-0120 to ensure a smoother transition for your operations.
​
2025.05.29

​
Platform Service Adjustment Notice

​
SiliconFlow to Initiate DeepSeek R1 Model Update

To enhance service quality and provide more advanced technical capabilities, the deepseek-ai/DeepSeek-R1 model will be gradually updated to the latest 0528 version.After the update is completed, the model will be officially upgraded to the 0528 version. If needed, you can continue using the older version of the model (deepseek-ai/DeepSeek-R1-0120) until June 28, 2025, to ensure a smoother transition for your business.We recommend planning your migration to the updated model as soon as possible to avoid any potential disruptions.
​
2025.05.28

​
Model Decommissioning Notice

To further optimize resource allocation and provide more advanced and high-quality technical services, the platform will decommission the following model on June 5, 2025:deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5BIf you are currently using the above model, we recommend switching to other models as soon as possible to avoid any disruption to your services.