Droid v0.53.0

Gemini 3 Flash Preview 2026-01-18
Score
13.44
FAILED
Pass Rate: 46.2%
go: 83.3% (6 tasks)
go
rust: 50.0% (6 tasks)
rus
typescript: 40.0% (5 tasks)
typ
kotlin: 33.3% (3 tasks)
kot
zig: 33.3% (3 tasks)
zig
dart: 0.0% (3 tasks)
dar

Evaluation Report

Summary

Metric Value
Agent droid
Model gemini-3-flash-preview
Reasoning Effort high
Timestamp 2026-01-18T164235
Pass Rate 46.2% (12/26)
Weighted Pass Rate 40.4%
Weighted Score 13.44 / 33.29
Duration 3699.2s

Quality Breakdown

  • Clean Passes: 10
  • Partial Passes (timed out but passed): 2
  • Integrity Violations (modified test files): 0
  • Failures: 12

Results by Language

Language Passed Failed Total Pass Rate
dart 0 3 3 0.0%
go 5 1 6 83.3%
kotlin 1 2 3 33.3%
rust 3 3 6 50.0%
typescript 2 3 5 40.0%
zig 1 2 3 33.3%

Results by Tier

Tier Passed Failed Total Pass Rate
core 5 7 12 41.7%
extended 7 7 14 50.0%

Task Results

Task Status Weight Score Duration
dart/future-pool ✘ FAIL 1.46 0.00 167.3s
dart/isolate-pool ✘ FAIL 1.50 0.00 196.7s
dart/reactive-cache ✘ FAIL 1.50 0.00 257.2s
go/bank-account ✔ PASS 1.04 1.04 26.6s
go/dining-philosophers ✔ PASS 1.04 1.04 97.4s
go/errgroup-limit ✘ FAIL 1.14 0.00 120.5s
go/parallel-letter-frequency ✔ PASS 1.04 1.04 51.7s
go/react ✔ PASS 1.14 1.14 146.2s
go/singleflight ✔ PASS 1.28 1.28 80.1s
kotlin/channel-multiplexer ✘ FAIL 1.50 0.00 223.3s
kotlin/flow-processor ✘ FAIL 1.50 0.00 246.6s
kotlin/lru-cache ✔ PASS 1.09 1.09 64.1s
rust/circular-buffer ✔ PASS 1.12 1.12 109.4s
rust/doubly-linked-list ✔ PASS 1.24 1.24 47.6s
rust/generational-arena ✔ PASS 1.24 1.24 155.9s
rust/macros ✘ FAIL 1.50 0.00 120.6s
rust/parallel-letter-frequency ✘ FAIL 1.12 0.00 121.3s
rust/regex-lite ✘ FAIL 1.40 0.00 120.7s
typescript/csv-lite ✘ FAIL 1.36 0.00 120.8s
typescript/forth ⚠ PARTIAL 1.26 0.95 120.8s
typescript/glob ✘ FAIL 1.14 0.00 120.8s
typescript/promise-pool ✘ FAIL 1.20 0.00 120.9s
typescript/react ✔ PASS 1.14 1.14 187.9s
zig/arena-allocator ⚠ PARTIAL 1.50 1.12 245.8s
zig/comptime-json ✘ FAIL 1.50 0.00 244.6s
zig/small-vector ✘ FAIL 1.34 0.00 184.4s

Verification

  • Harness Version: v1.3.0
  • Weight Version: 2.0
  • Tasks Hash: blake3:168696ded119da3ea973f15bde9c00151c127b3f54168196a84212827fa1b1f0
  • Results Hash: blake3:223394e89dcd3fd87f7fce6c266141ad35c5ee86d9f7484ae4167cf41a68089e

Generated by SanityHarness on 2026-01-18T164235

Run Statistics

Tasks Passed 12 / 26
Duration 62m
Cost (Est) -
By Language
dart 0%
go 83%
kotlin 33%
rust 50%
typescript 40%
zig 33%