Cline shipped v4.1.22 on GitHub, adding Bee (by HEOSSI) and Pareto Inference as providers and enabling Anthropic server-side refusal fallback so Claude models can fail over to another upstream provider on OpenRouter and Cline. It makes GPT-6.1 Sol the default model across ten OpenAI-compatible providers, refreshes the recommended list to Claude Sonnet 5.5 and Claude Opus 5.5, and fixes cancel, checkpoint, and token-counting bugs. Practitioners using non-pinned defaults should expect different models after update.
Cline shipped CLI v3.0.66 on GitHub, rebuilding binaries with Bun 1.4.2 so macOS 27 no longer kills them at launch and x64 builds run without AVX2. The release fixes Windows system-proxy hub startup, Esc-to-stop handling, Bedrock routing for GPT-6/GPT-5.6 and Nova models, and refreshes default models across many providers to GPT-6.1 Sol, Claude Sonnet 5.5, and others. Practitioners using unpinned models should expect changed defaults.
Cline shipped CLI v3.0.66 on GitHub, rebuilding binaries with Bun 1.4.2 so macOS 27 no longer kills them at launch and x64 builds run without AVX2. The release fixes Windows system-proxy hub startup, Esc-to-stop handling, Bedrock routing for GPT-6/GPT-5.6 and Nova models, and refreshes default models across many providers to GPT-6.1 Sol, Claude Sonnet 5.5, and others. Practitioners using unpinned models should expect changed defaults.
Cline shipped Desktop v0.0.38 on GitHub, adding Linux ARM64 packages, cloud session persistence, and a range of reliability fixes. The release also refreshes the default model catalog across providers, making GPT-6.1 Sol the default for OpenAI and most gateways.
Cline shipped Desktop v0.0.38 on GitHub, adding Linux ARM64 packages, cloud session persistence, and a range of reliability fixes. The release also refreshes the default model catalog across providers, making GPT-6.1 Sol the default for OpenAI and most gateways.
Anthropic released Claude Sonnet 5.5, the second model in the Claude 5.5 family, positioning it as a faster, lower-cost complement to Opus 5.5 for well-scoped coding and knowledge work. The company says it runs over 30% faster than Sonnet 5 and effectively costs up to 30% less per task despite identical per-token pricing, and it is the first Sonnet subject to the cybersecurity safeguards applied to Anthropic's most capable models. It is available everywhere, including Amazon Bedrock, and Haiku 5.5 is expected in the coming weeks.
Anthropic released Claude Sonnet 5.5, the second model in the Claude 5.5 family, positioning it as a faster, lower-cost complement to Opus 5.5 for well-scoped coding and knowledge work. The company says it runs over 30% faster than Sonnet 5 and effectively costs up to 30% less per task despite identical per-token pricing, and it is the first Sonnet subject to the cybersecurity safeguards applied to Anthropic's most capable models. It is available everywhere, including Amazon Bedrock, and Haiku 5.5 is expected in the coming weeks.
Cline released SDK v0.0.88 with reliability fixes for agent runtime Stop handling, session resumption, and content-filter reporting, plus a refreshed model catalog. Direct Anthropic requests gain server-side refusal fallbacks, and OpenRouter/Cline requests set provider.allow_fallbacks for upstream failover.
Cline released SDK v0.0.88 with reliability fixes for agent runtime Stop handling, session resumption, and content-filter reporting, plus a refreshed model catalog. Direct Anthropic requests gain server-side refusal fallbacks, and OpenRouter/Cline requests set provider.allow_fallbacks for upstream failover.
Cline shipped Desktop v0.0.41 on GitHub, bringing cloud sessions to the same models as Cline on the web (Cline Usage-Billing, ClinePass, ClineFree), plus UX fixes for drafts, session highlighting, and connector management. It also refreshed default models across providers including DigitalOcean, GMI Cloud, NanoGPT, Nvidia, and Ofox.
Cline shipped Desktop v0.0.41 on GitHub, bringing cloud sessions to the same models as Cline on the web (Cline Usage-Billing, ClinePass, ClineFree), plus UX fixes for drafts, session highlighting, and connector management. It also refreshed default models across providers including DigitalOcean, GMI Cloud, NanoGPT, Nvidia, and Ofox.