Anthropic

Tier S · Ship-to-prod default.

Claude Fable 5.2

Anthropic’s current Fable flagship for difficult planning and coding work, with a complete published Superbench 2.1 run set.

Try the active visual tests

Open the generated scene to test it, or compare that prompt with another model.

6 runs

Concourse Divide

Run date: 2026-11-10

Generation provenance: 2.1 · single-prompt

Superbench 2.1 brief · tool-free single-message authoring, no repair turn

Visible agent rerouting, collision avoidance, and recovery from blocked connectivity.

Ember Glider

Run date: 2026-11-10

Generation provenance: 2.1 · repository-run

Superbench 2.1 returning brief · original 1.x run protocol

Sunset gliding journey testing flight energy management, checkpoint flow, and atmospheric scene craft.

Lightkeeper's Atelier

Run date: 2026-11-10

Generation provenance: 2.1 · single-prompt

Superbench 2.1 brief · tool-free single-message authoring, no repair turn

Material differentiation, dynamic light, geometric shadows, and art direction.

Low-Poly Tower Defense

Run date: 2026-11-10

Generation provenance: 2.1 · repository-run

Superbench 2.1 returning brief · original 1.x run protocol

Diorama tower defense testing economy balance, wave design, placement rules, and combat readability.

Wrecklight Salvage

Run date: 2026-11-10

Generation provenance: 2.1 · single-prompt

Superbench 2.1 brief · tool-free single-message authoring, no repair turn

Three-phase game progression with persistent physical cargo and a visible late-stage consequence.

Archived model versions

The latest Claude Fable 5.2 runs are shown above. Earlier versions are preserved separately and are not counted in this current run set.

Editorial verdict

Where this model fits

Back to all benchmarks →

Claude Fable 5.2 replaces Fable 5.1 in the editorial S tier. Six Superbench 2.1 artifacts are published: the four new contracts under the tool-free single-prompt protocol, and the two returning 3D games under their original repository-run protocol. Delivery was through Claude Code CLI via Claude Max OAuth; modelUsage recorded the requested claude-fable-5-2 identity, with no independent server attestation. S tier is an editorial routing decision for the current Fable option, not a numeric leaderboard score or a claim that every brief passed a rubric.

Best for

  • Difficult planning and architecture
  • Agentic coding and automation workflows
  • Long-horizon coding sessions
  • Inspectable prompt-to-prototype work

Watch out

  • Independent interactive browser review of the published artifacts was not completed at publication; inspect each run
  • CLI delivery has no independent server attestation of the model identity
  • No public API pricing is recorded in the Superbash catalog
  • Local artifacts are not provider benchmark scores

Why it is ranked here

  1. Fable 5.2 is the direct successor to Fable 5.1, which previously held the Fable S-tier slot. Fable 5.1 remains published as an archived model version.
  2. A complete Superbench 2.1 set of six inspectable artifacts is published, covering the four new contracts and both returning 3D games.
  3. S tier is an editorial placement for demanding Fable-family work, not a claim that Fable 5.2 wins every benchmark or that these runs were scored.

Evidence

2026-11-10

Superbash editorial model ranking

Takeaway: Claude Fable 5.2 is currently placed in Tier S.

The November 2026 editorial roster places Claude Fable 5.2 at rank 3.

Open source →
2026-11-10

Claude Fable 5.2 Superbench 2.1 run record

Takeaway: Six published artifacts: four single-prompt Superbench 2.1 contracts and two returning 3D games under the original repository-run protocol.

The run record lists protocol, timings, token counts, and artifact paths. It is evidence of this site’s visual suite, not a provider benchmark or aggregate score.

Open source →

Local benchmark run

All six Superbench 2.1 briefs: four new contracts plus two returning 3D games

Superbash Learn sourceNovember 10, 2026Run record and published artifacts →
Sources, pricing details and methodology

Run identity

Model specifications

Codex platform ↗
Status
Available in this Codex session
Codex model ID
claude-fable-5-2
Run date
November 10, 2026
Access route
Claude Code via Claude Max OAuth; selected model identity recorded per run. No API pricing or capacity claims.

Benchmark sources

Skyhook Cargo, Concourse Divide, Lightkeeper’s Atelier and Wrecklight Salvage were each written in a tool-free single-prompt Claude Code session that received only the frozen Superbench 2.1 prompt. Ember Glider and Low-Poly Tower Defense ran under their original 1.x repository-run protocol, with tools allowed. Delivery was Claude Code CLI via Claude Max OAuth; modelUsage recorded the requested claude-fable-5-2 identity. There is no independent server attestation.

Claude Fable 5.2 Superbench 2.1 run record

These are inspectable local artifacts, not provider benchmark scores, passed rubrics or an aggregate rating. Per-artifact reports record static document checks completed at publication. Independent interactive browser review of gameplay was not completed; no brief is claimed as a pass or a fail.

Evaluation settings

    Local visual-run methodology and validation records

    Edition coverage

    Recorded means a published artifact, not a behavioral pass. Historical returning results retain their original protocol.

    Superbench 2.1: 6/6 recorded · 0 pending · 0 not run
    • 15 Ember Glider · repository-run · Recorded
    • 14 Low-Poly Tower Defense · repository-run · Recorded
    • 21 Skyhook Cargo · single-prompt · Recorded
    • 22 Concourse Divide · single-prompt · Recorded
    • 23 Lightkeeper's Atelier · single-prompt · Recorded
    • 24 Wrecklight Salvage · single-prompt · Recorded