Lava lamp v1 · Claude Code · run 499f4389 · 28.10.2026 09:39 Tallinn

Claude Opus 5.5

Extra high effort

Works

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

Compare →Transcript Source

Desktop capture of Claude Code (Opus 5.5)'s output on Lava lamp v1, xhigh effort, 28.10.2026 09:39 (run 499f4389) — **Desktop** · 621.7 KB · b2d701966ae3…

Phone capture of Claude Code (Opus 5.5)'s output on Lava lamp v1, xhigh effort, 28.10.2026 09:39 (run 499f4389) — **Phone** · 187.5 KB · 840827c91c0b…

Motion capture of Claude Code (Opus 5.5)'s output on Lava lamp v1, xhigh effort, 28.10.2026 09:39 (run 499f4389) — **Motion** · 623.0 KB · ad0f2863f995…

Build time
28m 09s
11th fastest of 12 models at extra high effort. Wall time, start to graded: 29m 07s.
Output tokens
151,558
11th fewest of 12 models at extra high effort.
Output speed
90 tok/s
6th highest of 12 models at extra high effort.
API cost
$6.40
The CLI’s list-price estimate, not a charge.

Along the row

Claude Opus 5.5 at every effort.

Down the column

Other models at extra high effort.

Fastest median first · this run: 28m 09s

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 3dd1b73f3b9e494d96a24a91f7a3f64739c572ee20041e16c3e95218da5a2ce7, 1301.7 KB.

Run ID: 499f4389-6016-4441-9913-86e675bbe470
Benchmark: lava-lamp v1
Experiment: Single run
Started: 28.10.2026 09:39 Tallinn
Graded: revision 1 · 28.10.2026 10:08

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
Requirement	Check	Blocker	Result
none (v1)	Self-contained index.html is present	Blocker	pass
none (v1)	Visible page content	Blocker	pass
none (v1)	Visible autonomous animation	Blocker	pass
none (v1)	Accessible on/off control	Extra	pass
none (v1)	390 px mobile layout has no horizontal overflow	Extra	pass
none (v1)	Local animation frame performance	Extra	pass
none (v1)	No browser errors	Blocker	pass
none (v1)	No attempted external or linked-resource requests	Blocker	pass

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
Field	Requested	Detected	Evidence
Model	claude-opus-5-5	claude-opus-5-5	harness result, 68 of 68 API requests
Effort	xhigh	xhigh	68 of 68 API requests
Harness	not recorded	2.1.312 (Claude Code)	Version as the CLI reported it; binary SHA-256 6b5ec691195e…
Deadline	60 min	29m 07s used	Completed

Tool decisions

116 calls · 113 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.

Bash44 calls · 41 accepted · 3 rejected · 25 failed
Edit44 calls · 44 accepted
Read17 calls · 17 accepted
Write8 calls · 8 accepted
Grep2 calls · 2 accepted · 1 failed
Glob1 call · 1 accepted

Telemetry

Time and usage.

Generation time: 28m 09s
First content: 1.9 s
Output speed: 89.8 tokens/s, whole run including tools
API-equivalent cost: $6.40 (list-price estimate, not a charge)
Actual charge: not reported
Billing: api

Input tokens

10,445,814 tokens

Cache reads10,178,18697%
Cache writes267,4922.6%
Uncached input136<0.1%

Output tokens

151,558 tokens

Reasoning94,44862%
Other output57,11038%

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: 244.4 KB, 56 redactions. Download transcript.txt

Source files

1 file · 29.1 KB

Files in the output bundle
File	Size	SHA-256	Redactions
index.html	29.1 KB	4167aa435120ae27ebcc3db3919ccf7d0e05694f13b221b7dc13c9f3c14d6501	0

Bundle SHA-256 4167aa435120ae27ebcc3db3919ccf7d0e05694f13b221b7dc13c9f3c14d6501

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
Revision	Graded	Works
1 (current)	28.10.2026 10:08	works

Redaction: automated scan run; checked by Jaanus on 28.10.2026 17:51. Redacted: 40 local-path, 3 user-path, 13 username.

Download run.json