dex-cube-turn-mujoco-stage-2-three-moves
SHA-256 a0839945ff634a4929e07b52619da37fe48eb30e4415b0ab3ef18c447b3b775c
Exact prompt file path was not published.
Versioned benchmark collection
Rubik's cube, three turns
Public ledger · prompt hash recorded · 9 measured runs across 1 target. Each entry is a single recorded attempt; repeat-run reliability is not established.
9Measured runs
1Frozen targets
9Model families
9Configurations
Collection scope
Recorded evidence, without a synthetic score.
Runs share the same frozen target and prompt within each test, with disclosed provider-specific stacks. Costs are recorded estimates as labelled in each ledger, not subscription invoices.
Targets
- Dex Cube — 3 Move
Model coverage
- Gemini 3.8 Flash · High
- GLM 5.3 Flash · Highest
- GPT-6 Astra · Max
- GPT-6 Sol · Max
- Grok 4.6 · xhigh
- Grok 4.7 · xhigh
- MiMo V2.6 Pro · Highest
- Claude Opus 5 · Max
- Claude Opus 5.5 · Max
Evidence available
- 9 runs on prompt a0839945ff63
- Prompt SHA-256 a0839945ff634a4929e07b52619da37fe48eb30e4415b0ab3ef18c447b3b775c
- Exact prompt file path was not published
Still missing
- No released Builder project, harness package, or skill download.
- Exact prompt file path was not published with this test.
- 5 related records are listed separately and are not attached to this prompt.
- Repeat-run reliability evidence is not established.
Public receipts
9 measured run ledgers
Open any row for its recorded workflow cost, timing, token usage, stack disclosure, artifact integrity, and explicit evidence gaps.
01Gemini 3.8 FlashHigh reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · pass · Prompt version v1$11.30 API-equivalent list price, not a marginal subscription cash chargerecorded estimate44m 30s wall-clockEnd-to-end workflowOpen 02GLM 5.3 FlashHighest reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · pass · Prompt version v1$1.44 Provider-recorded usage estimate, not an itemized subscription cash chargerecorded estimate1h 31m 15.1s wall-clockEnd-to-end workflowOpen 03GPT-6 AstraMax reasoning · Dex Cube — 3 Movejudge_pass_independently_replayed · PASS · Prompt version v1$15.97 Standard API-equivalent estimate from the Codex wrapper receipt, not a subscription invoice: this was a ChatGPT-subscription run, $0 was charged and Codex emits no cost field.recorded estimate43m 04s wall-clockEnd-to-end workflowOpen 04GPT-6 SolMax reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · PASS · Prompt version v1$10.43 Standard API-equivalent estimate from the Codex wrapper receipt, not a subscription invoice: this was a ChatGPT-subscription run, $0 was charged and Codex emits no cost field.recorded estimate131m 02s wall-clockEnd-to-end workflowOpen 05Grok 4.6xhigh reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · PASS · Prompt version v1$4.55 Provider-recorded usage estimate, not an itemized subscription cash chargerecorded estimate46m 37s wall-clockEnd-to-end workflowOpen 06Grok 4.7xhigh reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · PASS · Prompt version v1$13.68 Provider-recorded usage estimate, not an itemized subscription cash chargerecorded estimate67m 57s wall-clockEnd-to-end workflowOpen 07MiMo V2.6 ProHighest reasoning · Dex Cube — 3 Movejudge_fail_independently_replayed · pass · Prompt version v1$0.39 Provider-recorded usage estimate, not an itemized subscription cash chargerecorded estimate91m 58.1s wall-clockEnd-to-end workflowOpen 08Claude Opus 5Max reasoning · Dex Cube — 3 Movejudge_pass_independently_replayed · PASS · Prompt version v1$22.55 API-equivalent estimate from session token counts at Anthropic list pricerecorded estimate1h 42m 42s (38.7 min to the self-stop + 64 min after the single resume) wall-clockEnd-to-end workflowOpen 09Claude Opus 5.5Max reasoning · Dex Cube — 3 Movejudge_pass_independently_replayed · PASS · Prompt version v1$4.75 recorded total_cost_usd from workspace FINAL.json (modelUsage.claude-opus-5-5, costBasis=list, Anthropic list price). Charged to the Claude subscription ($0 charged).recorded estimate32m 29s wall-clockEnd-to-end workflowOpen
Prompt provenance
Exact prompt receipts
A path is shown only when the published record includes the prompt hash that identifies it. Campaign evidence and projections keep the hash gap visible.
Separate records
Related records with unknown prompt provenance
These published records stay separate from the prompt receipts above. Unknown provenance is why they are listed here.
Limits
Evidence and material gaps
- Qwen 3.8 Flash and DeepSeek V4.1 Flash have no Rubik’s Cube three-move extension attempt.