AI Daily · Oct 4, 2026: Cline offers Ling 3.1 Flash free until October 13; Command Code adds free open models
Cline integrates the 560B MoE Ling 3.1 Flash with free access until October 13, while Command Code adds several free open models with Ling 3.1 Flash boosted to 300 requests per day. Measured tests show LiteLLM cuts time-to-first-byte on long prompts by 94%.
Showing the latest briefing (Oct 4, 2026). Every day is generated the next morning.
Pick a date
-
Cline adds Ling 3.1 Flash with free access until October 13
Cline has integrated the Ling 3.1 Flash model, a 560B total-parameter MoE with 25B active parameters, which is said to match open-weight frontier models like Kimi K3 and DeepSeek V4 Pro, and is free to use until October 13.
-
Command Code adds free open models with daily request limits
Command Code now offers free models including Space Bunny Alpha with max reasoning, Ling 3.1 Flash with an increased 300 requests per day, Ling 3.0 Flash Sante, and Laguna S 2.1.
-
LiteLLM cuts time to first byte by 94 percent on long prompts
A report via LiteLLM claims time to first byte dropped 94% for long prompts, with median latency on a 440,000-token conversation falling from 553 ms to 35 ms.
-
Labs cut open model prices 50% this week
Command Code CEO states labs cut prices by 50% this week due to open models, noting his company is one of the largest buyers of open model tokens.
-
Step 5 Preview ranks seventh on Vals leaderboard at $2.54 per task
StepFun's Step 5 Preview debuted on the Vals leaderboard, ranking 7th among open-weight models—just ahead of Qwen 3.8 Max—at a cost of $2.54 per task.