CLAUDE CODE MARKETPLACES
All issues

November 18, 2026

Fable 5.2 & Mythos 5.2 ship with 30% speed gains and massive coding leaps; Claude Meetings, Sheets, and Whiteboards launch; Anthropic prices at $70–$78 and lists Thursday Nov 19

Get the next issue

WeeklyFree

TL;DR

A dense week. Anthropic dropped Fable 5.2 and Mythos 5.2 on November 9 — 30% faster, dramatically better coding benchmarks (FrontierCode 2.0: 49.1% vs 32.4% for Fable 5.1), and $8/$40 per million input/output tokens. Claude Meetings launched November 10 as a beta for Team and Enterprise — Claude joins Google Meet calls, drafts plans in real time, creates whiteboards and task lists, and posts notes to claude.ai after the call. The same day brought Claude Sheets (build full spreadsheets from a prompt), Claude Whiteboards (shared visual planning boards), and Claude Data GA (formerly Dashboards, now with Snowflake/BigQuery/Salesforce connectors). On the IPO front, Anthropic filed its final S-1/A price amendment — 825 million shares at $70–$78 per share ($1.3–1.45T valuation) — CNBC sources say pricing is expected above that range, with the stock listing on Nasdaq under ANT on Thursday November 19. Claude Code shipped four releases.

Model Releases

Fable 5.2Mythos 5.2Model ReleaseAPIBenchmarks

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9. Both share the same underlying model; Mythos 5.2 adds extra safeguards for biology, cybersecurity, and LLM R&D (access requires the Life Sciences or Cyber Verification programs). Key improvements over 5.1: output generation is up to 30% faster; at Medium effort the model finishes tasks with fewer tokens; safeguards are as robust as 5.1 but produce fewer false positives. Benchmarks: FrontierCode 2.0 at xhigh effort 49.1% (vs 32.4% for Fable 5.1 and 45.5% for Opus 5.5); FrontierSWE v2 69.9% (vs 65.5% for GPT-6 Astra); Remote Labor Index 41.3% (vs 21.3% for Opus 5.5); RoboDojo Sim 64.9% (vs 27.6% for Fable 5.1); AutomationBench 51.4% (behind Gemini 4 Carbon at 53.9%); AA-Briefcase Elo 1785 (behind Opus 5.5 at 1800). Fable 5.2 can act as an Advisor to smaller models or as an Orchestrator delegating simpler tasks. Pricing: $8/M input, $40/M output; optional fast mode: $16/$80/M. Fable 5.2 is now the default model in Claude Code as of v2.1.321. Anthropic reports that over 90% of changes merged into its own core codebase are now written by Claude, and a critical production service migrated from Python to Rust now starts 26× faster.

Products

Claude MeetingsCoworkBetaTeamEnterpriseGoogle Meet

Claude Meetings launched November 10 in beta for Team and Enterprise plans, with Google Meet as the first supported platform. Claude joins the call as a visible participant, identifies itself in the meeting chat, and works on tasks while the conversation unfolds: it drafts the plan being discussed, maps out timelines on whiteboards, answers questions in the chat, and retrieves figures from connected work apps. After the call, meeting notes, action items, and created materials are available in claude.ai. Safety guardrails: Claude will not join a call that includes external guests unless an org admin has explicitly authorized it, and it automatically leaves if a guest joins mid-call. Support for additional platforms is planned.

Claude SheetsWhiteboardsClaude DataBetaGAProProductivity

Three tools shipped together on November 10, all available on Pro and paid plans. Claude Sheets (beta): builds working spreadsheets from a text description — budgets, trackers, project plans — complete with formulas, charts, dropdown menus, and conditional formatting; it highlights cells it changed and exports to Excel. Claude Whiteboards (beta): turns unorganized ideas into a visual, editable shared board; Claude can group similar ideas, identify themes, and convert the board into a document or presentation; supports continued work in tools like Miro and Lucid. Claude Data (GA): formerly called Claude Dashboards; connects to Snowflake, BigQuery, and Salesforce via supported connectors; creates dashboards, charts, and written summaries with change tracking and data stories; certified queries and live sharing remain in beta.

IPO Update

IPOANTNasdaqFinanceS-1Company

Anthropic filed its final amended S-1 on November 12, setting a price range of $70–$78 per share on 825 million shares, implying a market valuation of $1.3–1.45 trillion — down from the originally targeted $2T but up from the $1.1–1.25T floor discussed during tech-selloff weeks. Nine-month revenue through September 30 was $33.2B, up 1,206% year-over-year (versus $4.6B for all of 2025). Nine-month operating loss was $1.3B, improved sharply from $5.2B a year earlier. The September quarter had a $536M operating loss including a $5B charge for short-term compute commitments and $2B in non-cash charges from the donation matching program. Aggregate committed spend stands at $895.4B, with $69.8B due in the next 12 months. CNBC reported November 17 that sources expect Anthropic to price above the stated $70–$78 range given strong institutional demand. Pricing is set for Wednesday November 18; Nasdaq trading begins Thursday November 19 under ticker ANT. Goldman Sachs and Morgan Stanley are lead underwriters across 22 total banks.

Claude Code

Claude CodeReleaseSecurityDXFable 5.2Permissions

Four releases shipped this week. v2.1.321 (Nov 9): Claude Fable 5.2 becomes the new default model; fixes four permission-flow gaps — read-only Bash commands running outside sandbox, file-write path checks using the wrong folder, deny rules not applying to files written by cp or mv, and MCP OAuth sending tokens to another host after a redirect. v2.1.322 (Nov 10): adds fast mode on Vertex AI in global, us, and eu regions, falling back to standard speed if Vertex refuses the request. v2.1.323 (Nov 12): Code Mode — an experimental opt-in where Claude writes JavaScript to call its tools rather than using them directly; plugins can now declare a supported version range in plugin.json via engines["claude-code"] so out-of-range plugins won't load; claude plugin validate now type-checks TypeScript hooks modules; a startup warning flags Bash allow rules with a wildcard in a variable value; .docx files now supported for Artifact tool uploads; progress heartbeat every 30 seconds for long-running hooks in claude -p and SDK sessions. v2.1.325 (Nov 13): fixes a security issue in zstd-compressed response decoding; sandboxed Bash commands now follow the normal permission flow by default (set sandbox.autoAllowBashIfSandboxed: true to keep the old behavior); scheduled tasks and /loop now ask before running a skill that declares allowed-tools or hooks; fixes context-length errors on some third-party gateways and Foundry, session resume, subagent model selection.

Action Items

  • →Upgrade to Fable 5.2 in Claude Code (now the default model as of v2.1.321) — it beats Fable 5.1 by a wide margin on coding benchmarks and runs 30% faster at the same price
  • →Team and Enterprise admins: enable Claude Meetings in Admin → Features — Claude can join Google Meet calls, draft plans and whiteboards in real time, and post notes to claude.ai; it won't join calls with external guests unless you authorize it
  • →Try Claude Sheets (beta) for spreadsheet generation from a description — supports formulas, charts, dropdowns, and Excel export; available on Pro, Max, Team, and Enterprise
  • →Claude Data is now GA (formerly Dashboards) — connect Snowflake, BigQuery, or Salesforce in Connectors and ask Claude to build dashboards and summaries; certified queries and live sharing are still beta
  • →Claude Code users: update to v2.1.325 — fixes a zstd decompression security issue and four permission-flow gaps where prompts were skipped; no config changes needed

Sponsor

This Week in Claude

Reach 3,200 Claude Code builders every Monday morning. Sponsors get a dedicated section, logo placement, and a link in the resources list. First come, first served — slots fill 3 weeks out.

View sponsorship details

If you found this useful, subscribe to get the next issue.

Get the next issue

WeeklyFree