Claude Code v2.1.170: Claude Fable 5 Model Access and VS Code Session Fix
Claude Code v2.1.170 introduces Claude Fable 5, a new Mythos-class model that Anthropic has deemed safe for general use. According to the release, Fable's capabilities exceed those of any model previously made generally available. The update also fixes a session-saving bug affecting VS Code integrated terminal users.
Key Changes
- Claude Fable 5 (Mythos-class): New high-capability model, safer for general deployment. Capabilities surpass all prior generally available models. Requires updating to v2.1.170 for access.
- Session Transcript Fix: Sessions launched from the VS Code integrated terminal (or any shell inheriting Claude Code environment variables) were not saving transcripts and failed to appear in
--resumeoutput. This is now fixed.
Getting the Update
Update Claude Code to version 2.1.170 to access Claude Fable 5. If you're using VS Code's integrated terminal, the session fix is especially relevant — your transcripts will now persist correctly.
Who This Affects
All Claude Code users interested in evaluating new models, and anyone relying on --resume for session continuity, particularly those using VS Code or custom shell environments that propagate Claude Code env vars.
📖 Read the full source: GitHub Claude-Code
👀 See Also
OpenClaw Client Adds Cost Tracking and Per-Agent Spending Limits
New release adds spending caps per agent, live usage UI with circular progress bar, sub-agent management, skill toggling, and per-agent model selection.
Meta to capture employee mouse movements and keystrokes for AI training
Meta plans to start capturing employee mouse movements and keystrokes for AI training data, according to a Reuters report. The article has generated discussion on Hacker News with 33 points and 7 comments.
Reddit Discussion on Claude's Impact on MVP Development and Founder Pitfalls
A Reddit user discusses how Claude AI lowers technical barriers for building MVPs from $3k-$5k to DIY, but warns about increased competition and founders focusing too much on building versus marketing, PMF, and operations.
Benchmark Results: Qwen3.5 Models on Apple Silicon vs AMD GPUs with ROCm vs Vulkan
A developer benchmarked Qwen3.5 models (35B MoE, 27B dense, 122B MoE) across Apple Silicon Macs and AMD GPU workstations, comparing ROCm and Vulkan backends with context-scaling tests. Hardware included M5 Max, M1 Max, and three AMD GPUs with different PCIe configurations.