CmdBenchby CmdBrief
Menu

Lava lamp v1 · Codex CLI · run e375a13d · 08.10.2026 19:01 Tallinn

GPT-6 Astra

Low effort

Not verified

No checks ran on this run, so it has no verdict.

Compare →TranscriptSource

Along the row

GPT-6 Astra at every effort.

  1. No capture published Low4m 17s

    Not checked

    This effort
  2. GPT-6 Astra at medium effort: desktop capture of run d0b2c47f Medium4m 15s

    Works

    1 run
  3. GPT-6 Astra at high effort: desktop capture of run f24ff81c High7m 30s

    Works

    1 run
  4. GPT-6 Astra at extra high effort: desktop capture of run 47bb0692 Extra high25m 55s

    Works

    1 run
  5. GPT-6 Astra at max effort: desktop capture of run 9b8c977c Max36m 32s

    Works

    1 run

Down the column

Other models at low effort.

Fastest median first · this run: 4m 17s

  1. GPT-5.6 Luna at low effort: desktop capture of run 4b99cb20GPT-5.6 Luna29.7 s

    Works

    8.7× faster than this run
  2. GPT-6.1 Terra at low effort: desktop capture of run e58b2b0cGPT-6.1 Terra42.9 s

    Works

    6.0× faster than this run
  3. GPT-6 Luna at low effort: desktop capture of run 46ad720eGPT-6 Luna51.5 s

    Fails

    5.0× faster than this run
  4. GPT-5.6 Terra at low effort: desktop capture of run 0512c0d4GPT-5.6 Terra1m 38s

    Works

    2.6× faster than this run
  5. Claude Opus 5.5 at low effort: desktop capture of run dc76fa3bClaude Opus 5.51m 42s

    Works

    2.5× faster than this run
  6. GPT-5.6 Sol at low effort: desktop capture of run 184f5448GPT-5.6 Sol1m 46s

    Works

    2.4× faster than this run
  7. Claude Haiku 5.5 at low effort: desktop capture of run 69b2dfe6Claude Haiku 5.52m 36s

    Works

    1.6× faster than this run
  8. Claude Sonnet 5.5 at low effort: desktop capture of run b7f9c96bClaude Sonnet 5.52m 37s

    Works

    1.6× faster than this run
  9. GPT-6 Sol at low effort: desktop capture of run 0beb33c6GPT-6 Sol2m 39s

    Works

    1.6× faster than this run
  10. Claude Fable 5.1 at low effort: desktop capture of run 4bd42367Claude Fable 5.12m 50s

    Works

    1.5× faster than this run
  11. GPT-6.1 Sol at low effort: desktop capture of run 80d2eb3dGPT-6.1 Sol3m 11s

    Works

    1.3× faster than this run

Hidden checks

Not verified.

No checks are recorded for this run.

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

Run detailsagent config, what ran, tokens, source and revisions
Run ID
e375a13d-fb5e-4a84-9b41-10975347de6c
Benchmark
lava-lamp v1
Experiment
Single run
Started
08.10.2026 19:01 Tallinn
Graded
revision 1 · 08.10.2026 19:05

Works and checks, in full

No checks ran on this run, so it has no verdict.

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. It is never compared with contract v2 runs.

No checks are recorded for this run.

Verification state not recorded; eligibility not recorded.

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
Modelgpt-6-astranot reportedThe harness reported no model.
Effortlownot reportedThe harness reported no effort; the requested effort is shown.
Harnessnot recordedcodex-cli 0.161.0Version as the CLI reported it; binary SHA-256 12ac11d2c7ee…
Deadline10 min4m 17s usedCompleted

Tool decisions

3 calls · 3 without a decision

The harness reports no permission decision for its tool calls; the result of each call is still recorded.

  • command_execution2 calls · 1 failed
  • file_change1 call

Telemetry

Time and usage.

Generation time
4m 17s
First content
4.6 s
Output speed
30.1 tokens/s, whole run including tools
API-equivalent cost
not reported
Actual charge
not reported
Billing
subscription

Input tokens

62,049 tokens
  • Cache reads44,92872%
  • Cache writesnot reported
  • Uncached input17,12128%

Output tokens

7,727 tokens
  • Output7,727100%
  • Reasoningnot reported

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: 2.6 KB, 0 redactions. Download transcript.txt

Source files

1 file · 9.8 KB
Files in the output bundle
FileSizeSHA-256Redactions
index.html9.8 KB5362b6382169419093e453a5ad9b732b4353c8ac206b4c4125c8b74894da245e0
Bundle SHA-256 5362b6382169419093e453a5ad9b732b4353c8ac206b4c4125c8b74894da245e

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)08.10.2026 19:05pending

Redaction: automated scan run; checked by Jaanus on 22.10.2026 15:37. Nothing needed redacting.

Download run.json