Claude Fable 5.1

Fiche modèle · Anthropic · claude-fable-5-1

Claude Fable 5.1 is Anthropic's generally available, production-safeguarded deployment of the same underlying weights as Claude Mythos 5.1 (trusted-access only via CVP/LSVP; not cataloged here). It targets demanding reasoning and long-horizon agentic work with text and image input, text output, multilingual and vision support, tool use, adaptive thinking always on (API/Claude Code default effort high; Cowork/claude.ai default medium), a 1M-token context window, and 128K max output on the sync Messages API. Reliable knowledge and training-data cutoffs are June 2026; comparative latency is slower than the rest of the current lineup. First-party pricing matches Fable 5 at $10/$50 per million input/output tokens, with cache reads cut to $0.25 per million tokens (0.025x base input, down from $1 / 0.1x on Fable 5); Anthropic estimates ~25% lower cost on typical workloads and ~45% on highly agentic ones. Self-reported launch benchmarks with production safeguards enabled include Terminal-Bench-Science 0.1 52.6% (SE ±3.5–4.5 pts), Terminal-Bench 4.0 55.8%, GDPval-AA v2 1853 Elo, OSWorld 2.0 77.9% partial / 41.7% strict (August 2026 task release), Humanity's Last Exam 60.9% no tools / 65.0% with tools, AutomationBench 31.4%, and CursorBench 3.2.0 73.4%. Available on the Claude API as `claude-fable-5-1`.

Identité

OrganisationAnthropic
Pays—
Famille—
Typellm
Date de sortie2026-09-01
Licenceproprietary
Licence commercialenon
Licence open sourcenon
Poids ouvertsnon
Auto-hébergement (DGX Spark)non — poids fermés
Modalitésimage|text
Fenêtre de contexte1 000 000
KV cache / token (Q8)—
Paramètres totaux≈ 5 T (estimation, confiance très faible — https://www.reuters.com/technology/bytedance-targets-mega-ai-model-nearing-anthropics-mythos-ft-reports-2026-08-07)
Paramètres actifs (session)≈ 87 B–114 B (estimation, confiance très faible — https://www.reuters.com/technology/bytedance-targets-mega-ai-model-nearing-anthropics-mythos-ft-reports-2026-08-07)
Source des paramètres actifshttps://www.reuters.com/technology/bytedance-targets-mega-ai-model-nearing-anthropics-mythos-ft-reports-2026-08-07 (estimation, confiance très faible)
Curation des paramètres actifs2026-10-27
Prix par session20.00 $ (1 000 000 tokens entrée + 200 000 sortie, prix le moins cher par fournisseur)
Tokens / tâche (Data Players)1.2 M
Tokens / tâche (Terminal-Bench)40.9 M
Ratio Terminal-Bench / Data Players34.1×
Coût / tâche (Terminal-Bench)93.94 $
Tokens d'entraînement—
Coupure de connaissance—
Créé le2026-09-01T18:45:51.019009Z
Mis à jour le2026-11-04T06:04:32.506758Z
SourceLLM Stats
URLhttps://llm-stats.com/models/claude-fable-5-1

Prix par fournisseur

Statut du modèle (offres + auto-hébergement) : non.

Toutes sources confondues : catalogue llm-stats, routeurs EUrouter/Cortecs et catalogues directs (OVHcloud, Scaleway, Infomaniak). Les prix sont convertis en USD pour la comparaison ; le prix publié et le taux BCE utilisé sont dans l'infobulle. Les prix inconnus restent indiqués « — ».

FournisseurSourceSouverainetéEntrée (/M tok)Sortie (/M tok)QuantificationContexteLatence (s)Débit (tok/s)Statut
Anthropicllm-statsnon10.00 $50.00 $————active

Souveraineté qualifiée par offre (modèle × fournisseur) ; survolez le badge pour les critères.

Souveraineté

Scores par catégorie (top_scores)

CatégorieScore
legal1 853.00
math0.70
multimodal0.80
reasoning371.00
finance1 853.00
general309.30
agents371.00
code0.50
tool_calling0.50
vision0.70

Scores par benchmark

BenchmarkCatégorieScoreMaxNormalisé (API)VérifiéAuto-déclaréDaté du
automationbenchagents0.311.000.31nonoui2026-11-04T06:04:32.563151Z
cursorbench-3.2code0.731.000.73nonoui2026-11-04T06:04:32.573261Z
gdpval-aageneral1 853.003 000.000.62nonoui2026-11-04T06:04:32.523137Z
humanity's-last-examreasoning0.651.000.65nonoui2026-11-04T06:04:32.553154Z
lifesci-diamond-2026-11healthcare0.141.000.14nonnon2026-11-04T06:04:32.582388Z
osworld-2.0multimodal0.781.000.78nonoui2026-11-04T06:04:32.533087Z
terminal-bench-4.0reasoning0.561.000.56nonoui2026-11-04T06:04:32.517733Z
terminal-bench-science-0.1reasoning0.531.000.53nonoui2026-11-04T06:04:32.513426Z

Classements

CatégorieRangScoreNote conservatriceBenchmarks évaluésMéthodeDaté duProvenance
agents14371.0038.295trueskill2026-11-04T08:08:47.839150Zrankings
code360.5035.712trueskill2026-11-04T08:08:39.742311Zrankings
math20.7044.441trueskill2026-11-04T08:08:42.442591Zrankings
multimodal270.8032.081trueskill2026-11-04T08:08:46.488887Zrankings
reasoning14371.0050.075trueskill2026-11-04T08:08:41.093494Zrankings

Lecture : Note conservatrice = borne basse TrueSkill (niveau estimé − marge d’incertitude) ; Benchmarks évalués = nombre de benchmarks de la catégorie pris en compte. À performances égales, un modèle évalué sur moins de benchmarks obtient une note plus basse : la note ne se déduit pas d’un seul score. Provenance : rankings = top 50 de /rankings ; model_history = historique du modèle (/models/{id}/history), utilisé pour les rangs au-delà du top 50.

Données

Dates de curation (récupération depuis l'API) : models 2026-11-04 · benchmarks 2026-11-04 · scores 2026-11-04 · updates 2026-11-04 · account 2026-11-04 · model_history 2026-11-04 · rankings 2026-11-04