BuilderBuilder
GPT-6.1 Sol · Robot pick-and-place
The episode's pick-and-place chapter matches the unaided FR3 apple-carry max runs. All four judge-pass. The recorded API-equivalent totals are $2.68, $0.32, $2.00, and $1.61, all under $3. Astra and Sol display 8m 00s. Opus displays 25m 29s and Sonnet 24m 14s. The episode says Astra is not fully rendered, which matches Astra's pinned capture: it is not a cycles-hero file. Opus, Sonnet, and GPT-6.1 Sol have hash-matched cycles-hero renders. HERO rendered capture Unaided model Open this round
Measured harness ledgerPublic result
GPT-6.1 Sol

Robot pick-and-place — GPT-6.1 Sol

Carry an apple onto a plate on the FR3 with a Robotiq 2F-85. The episode chapter is this unaided coding-agent task, not apple-stem and not CAD-002.

Max reasoningHeadline result
Workflow cost
$0.32
Wall-clock
8m 00s
Processed tokens
964,455 processed (fresh input 57,490 + cache read 895,872 + cache creation 0 + output 11,093). Reasoning output 4,942 is a subset of output and is not added.
Record state
additive_harness_projection
Public summary

GPT-6.1 Sol Robot pick-and-place: 8m 00s, $0.32 API-equivalent estimate, 964,455 processed (fresh input 57,490 + cache read 895,872 + cache creation 0 + output 11,093). Reasoning output 4,942 is a subset of output and is not added..

Run identity and stack
  • Result ID: apple-carry-fr3-gripper-gpt-6.1-sol-max-codex-2026-10-02
  • Technical model: gpt-6.1-sol
  • Provider: OpenAI Codex
  • Client: Codex CLI 0.160.0
  • Stack: FR3 plus Robotiq embodiment
  • Stack: Harness apple-carry prompt
Cost basis
  • Wall clock is supplied_measurements.wall_clock_display. Cost is supplied_measurements.cost.total API-equivalent. Charged amount is $0.
  • Source: supplied_measurements.token_usage.total_processed at harness commit 9b7c614dbb5103cc373fdf6275b5c4efc10d8f0e.
Validation evidence
  • Result: PASS
  • Verification: Task verdict is judge_pass_independently_replayed. The replay gate is that status. Additive projection from harness commit 9b7c614dbb5103cc373fdf6275b5c4efc10d8f0e. Broad harness sync was not run. Wall clock is supplied_measurements.wall_clock_display. Cost is supplied_measurements.cost.total API-equivalent. Charged amount is $0. Source: supplied_measurements.token_usage.total_processed at harness commit 9b7c614dbb5103cc373fdf6275b5c4efc10d8f0e.
Recorded caveats
  • Judge pass, 5 of 5 seeds, stage 3, embodiment fr3-gripper. Cycles-hero sha256 matches. The hero does not change the judge result.
  • Wall clock is supplied_measurements.wall_clock_display. Cost is supplied_measurements.cost.total API-equivalent. Charged amount is $0.
  • Source: supplied_measurements.token_usage.total_processed at harness commit 9b7c614dbb5103cc373fdf6275b5c4efc10d8f0e.
Visible evidence gaps
  • additive projection; broad harness sync was not run
  • source harness commit 9b7c614dbb5103cc373fdf6275b5c4efc10d8f0e
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.

Browse BuilderReturn to GPT-6.1 Sol
RemakeBenchResearch console