Lava lamp v1 · Claude Code · run 45666ded · 09.11.2026 22:30 Tallinn

Claude Fable 5.2

High effort

Works

8 of 8 checks passed. All 5 blocker checks passed.

What it built: 3 captures

Desktop capture of Claude Code (Fable 5.2)'s output on Lava lamp v1, high effort, 09.11.2026 22:30 (run 45666ded)
Desktop · 565.2 KB · e141ffaffa7a…
Phone capture of Claude Code (Fable 5.2)'s output on Lava lamp v1, high effort, 09.11.2026 22:30 (run 45666ded)
Phone · 108.4 KB · b1356404970a…
Motion capture of Claude Code (Fable 5.2)'s output on Lava lamp v1, high effort, 09.11.2026 22:30 (run 45666ded)
Motion · 566.2 KB · d9ff828e8c3b…

Along the row

Claude Fable 5.2 at every effort.

  1. Claude Fable 5.2 at low effort: desktop capture of run 96e988f7 Low4m 43s

    Works

    1 run
  2. Claude Fable 5.2 at medium effort: desktop capture of run b79ff6d9 Medium7m 47s

    Works

    1 run
  3. Claude Fable 5.2 at high effort: desktop capture of run 45666ded High34m 55s

    Works

    This effort
  4. Claude Fable 5.2 at extra high effort: desktop capture of run e47eb2fe Extra high40m 03s

    Works

    1 run
  5. MaxNot run

Down the column

Other models at high effort.

Fastest median first · this run: 34m 55s

  1. Muse Ember at high effort: desktop capture of run 9f5b66e4Muse Ember28.0 s

    Works

    74.8× faster than this run
  2. GPT-5.6 Terra at high effort: desktop capture of run 7802fc2cGPT-5.6 Terra1m 00s

    Works

    34.9× faster than this run
  3. GPT-5.6 Luna at high effort: desktop capture of run e3bceba4GPT-5.6 Luna1m 04s

    Works

    32.5× faster than this run
  4. GPT-6.1 Terra at high effort: desktop capture of run 647bf11fGPT-6.1 Terra1m 12s

    Works

    29.1× faster than this run
  5. GPT-6 Luna at high effort: desktop capture of run c6616e3dGPT-6 Luna1m 27s

    Works

    24.0× faster than this run
  6. GPT-5.6 Sol at high effort: desktop capture of run 456bc2ffGPT-5.6 Sol2m 34s

    Works

    13.6× faster than this run
  7. GPT-6 Sol at high effort: desktop capture of run 5f2b6af6GPT-6 Sol4m 12s

    Works

    8.3× faster than this run
  8. GPT-6.1 Sol at high effort: desktop capture of run fa862c30GPT-6.1 Sol5m 15s

    Works

    6.7× faster than this run
  9. Claude Haiku 5.5 at high effort: desktop capture of run c23d1490Claude Haiku 5.55m 22s

    Works

    6.5× faster than this run
  10. Claude Sonnet 5.5 at high effort: desktop capture of run 4c8fd1eaClaude Sonnet 5.56m 45s

    Works

    5.2× faster than this run
  11. GPT-6 Astra at high effort: desktop capture of run f24ff81cGPT-6 Astra7m 30s

    Works

    4.7× faster than this run
  12. GPT-6.2 Astra at high effort: desktop capture of run 6b64fc54GPT-6.2 Astra9m 43s

    Works

    3.6× faster than this run
  13. GPT-6.2 Sol at high effort: desktop capture of run cc6c59e3GPT-6.2 Sol11m 10s

    Works

    3.1× faster than this run
  14. Claude Opus 5.5 at high effort: desktop capture of run 00eeb954Claude Opus 5.513m 48s

    Works

    2.5× faster than this run
  15. Claude Fable 5.1 at high effort: desktop capture of run 26f3f534Claude Fable 5.114m 06s

    Works

    2.5× faster than this run

Hidden checks

8 of 8 checks passed.

  • Self-contained index.html is presentblocker
  • Visible page contentblocker
  • Visible autonomous animationblocker
  • Accessible on/off controlextra
  • 390 px mobile layout has no horizontal overflowextra
  • Local animation frame performanceextra
  • No browser errorsblocker
  • No attempted external or linked-resource requestsblocker

It works when every blocker passes. How it looks is reviewed blind: pending.

Run detailsagent config, what ran, tokens, source and revisions
  • Video (video): not published, videos are withheld. SHA-256 72a12eaaaeaa0ab95d067d39c7b8cc3550e4609a5a73658844c6aeb69fb015c2, 1560.1 KB.
Run ID
45666ded-0cc1-4307-aef5-3602acca2b2b
Benchmark
lava-lamp v1
Experiment
Single run
Started
09.11.2026 22:30 Tallinn
Graded
revision 1 · 09.11.2026 23:06

Works and checks, in full

All 5 blocker checks passed.

The rule: a run works when every blocker check passes. If any blocker fails it does not work. If none fails but one has no result yet, it stays pending, never counted as a pass or a fail. Extra checks are reported but do not decide it. Works is not a judgement of looks: that is the visual review.

This is a legacy contract v1 run: the task has no public requirements, and these are the original lava-lamp checks, not a versioned hidden verifier. It is never compared with contract v2 runs.

Contract v1 checks for this run, with blocker and result
RequirementCheckBlockerResult
none (v1)Self-contained index.html is presentBlockerpass
none (v1)Visible page contentBlockerpass
none (v1)Visible autonomous animationBlockerpass
none (v1)Accessible on/off controlExtrapass
none (v1)390 px mobile layout has no horizontal overflowExtrapass
none (v1)Local animation frame performanceExtrapass
none (v1)No browser errorsBlockerpass
none (v1)No attempted external or linked-resource requestsBlockerpass

Verification completed; eligible for benchmark results.

Visual review

Pending.

Pending: the captures have not been reviewed blind yet. Checks decide whether it works; how it looks is decided only in that review.

What ran

Requested and detected.

Detected values come from the run’s own output, never from our settings.

Requested agent config beside what the run reported
FieldRequestedDetectedEvidence
Modelclaude-fable-5-2claude-fable-5-2harness result, 49 of 49 API requests
Efforthighhigh49 of 49 API requests
Harnessnot recorded2.1.321 (Claude Code)Version as the CLI reported it; binary SHA-256 abc9feeeca80…
Deadline60 min35m 53s usedCompleted

Tool decisions

95 calls · 92 accepted · 3 rejected

Each tool call is allowed or blocked by the harness’s permission settings before it runs. Allowed is not the same as succeeded: an allowed call can still fail.

  • Bash39 calls · 36 accepted · 3 rejected · 39 failed
  • Edit30 calls · 30 accepted
  • Read17 calls · 17 accepted
  • Write9 calls · 9 accepted

Telemetry

Time and usage.

Generation time
34m 55s
First content
0.9 s
Output speed
81.4 tokens/s, whole run including tools
API-equivalent cost
$11.19 (computed from the run’s token counts at list prices, not a charge)
Actual charge
not reported

Input tokens

5,254,953 tokens
  • Cache reads5,044,56696%
  • Cache writes210,2874.0%
  • Uncached input100<0.1%

Output tokens

170,438 tokens
  • Reasoning97,00257%
  • Other output73,43643%

Usage as the agent reported it at the end of the run. Cache reads are part of input; reasoning is part of output.

Transcript and source

Plain-text downloads.

The redacted transcript and the frozen output bundle, with LF line endings. Hashes are of the original files.

Redacted agent transcript: 206.4 KB, 53 redactions. Download transcript.txt

Source files

1 file · 44.4 KB
Files in the output bundle
FileSizeSHA-256Redactions
index.html44.4 KB7950a476ce0132b7a58a64ed14f6c7bbc312ebc24c36050ff4de4e3007e06c7a0
Bundle SHA-256 7950a476ce0132b7a58a64ed14f6c7bbc312ebc24c36050ff4de4e3007e06c7a

Revision history

This page shows revision 1.

Grades are append-only: a regrade adds a revision and never rewrites an earlier one.

Grading revisions of this run
RevisionGradedWorks
1 (current)09.11.2026 23:06works

Redaction: automated scan run; checked by Jaanus on 10.11.2026 07:10. Redacted: 34 local-path, 6 user-path, 13 username.