Measured harness ledgerPublic result
MiMo V2.6 Pro

JRPG boss battle — MiMo V2.6 Pro High · Attempt 2

Build an interactive Three.js JRPG boss battle with supplied character assets, required combat beats, animation, sound controls, effects, and a validator-ready game loop.

High reasoningHeadline result
Workflow cost
$0.31
Wall-clock
1782.6s wall-clock
Processed tokens
3.18M processed
Record state
partial
Public summary

MiMo V2.6 Pro High partial ledger: 1782.6s wall-clock, 3.18M processed, and $0.31 live OpenCode OpenRouter model entry and exact one-turn session usage ledger.

Run identity and stack
  • Result ID: jrpg-boss-battle-mimo-v2.6-pro-high-attempt-2
  • Technical model: MiMo-v2.6-Pro
  • Provider: OpenRouter
  • Client: OpenCode 1.18.23
  • Attempt: 2
  • Stack: OpenRouter
  • Stack: OpenCode 1.18.23
  • Stack: OpenCode → OpenRouter
  • Stack: Technical model/configuration: MiMo-v2.6-Pro
  • Stack: Attempt 2
  • Stack: Three.js / Vite game
  • Stack: Supplied GLB fixture bundle
  • Stack: Harness v1 JRPG validator
  • Stack: Requested tool profile: OpenCode/OpenRouter plus task-prescribed bundled Codex CLI image generation; run-local Blender MCP disabled
Primary artifact integrity
  • Kind: manifest-verified-artifact
  • Path: artifacts/jrpg-boss-battle-mimo-v2.6-pro-high-attempt-2/source/src/main.js
  • SHA-256: 0428b9b3f1b8b3cd367af20364d45e6bce5d691638e42c5e47d04909d215817d
Recorded caveats
  • Model internal Done is preserved verbatim, but it is not accepted as proof of an ordinary playable win path.
  • The model's NOTES.md and final response overstate the victory playthrough: its own script wrote bossHp before the finishing keyboard attack.
  • One independent input-only natural battle ended in defeat; the documented damage bounds show a natural win is unreachable in the frozen source.
  • Wall clock includes tool time, nested image generation, validation, and idle time, not model-only compute.
  • Private OpenCode session/ledger contents remain local; aggregate public metrics and SHA-256 commitments are published.
Public result only

This result keeps its public summary and evidence, but it does not currently have a matching Builder test with prompts, projects, Harness workflows or skills.

RemakeBenchResearch console