Claude Fable 5.2
anthropic/claude-fable-5.2claude-fable-5-2claude-fable-5.2fable-5-2fable-5.2- 📏 One-million-token context window in Anthropic's Fable tier
- 🧠 Reasoning, vision, coding, function calling and web search
- 🔧 Built for agentic, long-horizon coding and tool-using workflows
- 👁️ Vision input alongside long-document and large-codebase reasoning
- 🆕 Successor to Fable 5.1 in Anthropic's top Claude tier
- 🏢 Mythos 5.2 variant mirrors its specs for verified organizations
- 🔒 Requires 30-day data retention; zero-retention only if authorized
- 🗓️ Released November 9, 2026 by Anthropic
Anthropic PBC is an American artificial intelligence company headquartered in San Francisco. Structured as a public benefit corporation, the lab develops large language models under the Claude name, with a research emphasis on building reliable, steerable, and safety-focused AI…
Explore 18 more models by Anthropic →Claude Fable 5.2 is Anthropic's November 2026 release in the Fable line, the company's top-end Claude tier for demanding reasoning and long-horizon agentic work. It carries forward the family's defining specification: a one-million-token context window, letting it ingest entire codebases or book-length document sets in a single pass. Catalog capabilities span reasoning, vision, code-oriented work, function calling and web search.
The Fable line began with Claude Fable 5 in mid-2026, documented by Anthropic as a model for complex knowledge work and coding with sustained autonomous operation, planning across stages, delegation to sub-agents and self-verification of its work. Claude Fable 5.1 followed in the same tier later that year. Anthropic's own Claude Code documentation positions Fable models as the most capable options available there, suited to tasks larger than a single sitting, with longer autonomous sessions and more verification than the smaller Claude models.
Fable 5.2 continues that trajectory as the current head of the family. As a covered model it requires 30-day data retention and is not offered under zero data retention without explicit authorization, and a parallel Mythos 5.2 release shares its specifications for organizations cleared through Anthropic's verification programs. Within the broader Claude range, Anthropic's Claude Code documentation describes Fable models as intended for larger-scope, longer-running tasks than the Sonnet and Haiku options. Readers evaluating it against siblings such as Claude Opus 5.5, Claude Sonnet 5.5 or Claude Haiku 5.5 should consult Anthropic's model documentation for current per-model limits.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · Claude Fable 5.2 · Organic sellers
Top sellers · Claude Fable 5.2 · Organic
- ▲ Apex Ant52.5%
- Venice.ai Proxy36.3%
- Super Seeder9.2%
- ZLKPro-Api1.7%
- bartly.eth64.de0.1%
- Fire Ant 🔥🐜0.1%
- Others0.0%
| Seller | Health | Rep↓ | Sales | Tokens | Routing | Input $/M | Cached $/M | Output $/M | Save |
|---|---|---|---|---|---|---|---|---|---|
| Super Seeder 0xd19f…41f3 | healthy | 89.22 | $2.68 | 773.97k | #4 | $10.0002 | $0.25 | $50.0009 | — |
| 💬 Banter 0x4122…e194 | healthy | 82.88 | $0.00 | 201.00 | #3 | $5.56 | $0.139 | $27.80 | −31% |
| ▲ Apex Ant 0x73b4…e736 | — | 81.71 | $15.28 | 22.32M | #1 | $4.00 | $0.10 | $20.00 | −50% |
| ZLKPro-Api 0x0b0b…f446 | healthy | 20.91 | $0.50 | 817.90k | gated | $0.85 | $0.40 | $4.50 | −89% |
| Unified Rack · DeepSeek 240 tok/s 0x42d1…822f | healthy | 13.21 | — | — | gated | $7.20 | $0.18 | $36.00 | −10% |
| OpenCheap 0x0ee1…6c41 | — | 12.43 | — | — | gated | $4.312 | $0.1078 | $21.56 | −46% |
| bartly.eth64.de 0x666e…4666 | healthy | 7.22 | $0.03 | 54.66k | gated | $0.30 | $0.015 | $1.50 | −96% |
| LLMart 0x948a…a000 | — | 0.00 | $0.00 | 202.00 | gated | $5.55 | $0.1426 | $27.79 | −31% |
| Fire Ant 🔥🐜 0xbe05…bc5d | — | 0.00 | $0.03 | 38.07k | gated | $5.55 | $0.129 | $27.79 | −31% |
| Hana Gateway ✅ 0x4ae1…117b | healthy | 0.00 | — | — | gated | $1.00 | $0.20 | $5.00 | −88% |
| Venice.ai Proxy 0x1f22…18c9 | failingupstream error · 3.7 h | 86.84 | $10.57 | 24.06M | #2 | $4.40 | $0.11 | $22.00 | −45% |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.