DeepSeek Evolved While You Waited
The anticipated DeepSeek has already transformed significantly. The article recaps key developments over the past year.
Search
Tag: #model-update28 results
Nano Banana 2 Faster Image Generation
Google's Nano Banana 2 improves on predecessor with faster speeds, better text rendering, and higher resolutions. It's the new default image generation model. Guide provided on how to try it.
New Kimi Model Variant Released on Hugging Face
A new model variant, novita/kimi-k2.6-dspark, has been uploaded to Hugging Face. This update likely pertains to the Kimi series of models developed by Moonshot AI.
🤖
Claude Opus 4.7 Token Consumption Surges
Claude Opus model version 4.7 exhibits significantly higher token consumption according to a tech news roundup. This change likely raises costs for API users processing similar workloads. No further technical details provided in the brief.
🤖
Gemini Chat Histories Vanish Post-3.1 Update
Users report months of Google Gemini conversations disappearing recently. Issues coincide with Gemini 3.1 rollout; Google attributes to temporary bug.
Minimax 2.7 Update Incoming
Community excitement builds for the Minimax 2.7 model update, clarified by Yuanhe134. LocalLLaMA subreddit users eagerly await this important release. Posted 2 hours ago by /u/LegacyRemaster.
Qwen-Code Nightly v0.13.2 Released
Qwen released the nightly build v0.13.2-nightly.20260331.1b1a029fd for qwen-code on GitHub. The full changelog details changes from v0.13.2. Developers can access it via Qwen's GitHub Releases.
Qwen Code v0.12.5 Nightly Release
Qwen released nightly build v0.12.5-nightly.20260317.dbfa5b3e8 for qwen-code on GitHub. Full changelog covers changes from v0.12.5 base version. Developers can access latest tweaks promptly.