Touchstone
← Benches

ladder.csharp.title_case

C#gauntlet

An owned ladder bench, graded mechanically against a fixed test.

Solved
22/24
runs passed
Models
23
have attempted
Harnesses
2
scaffolds tried

Runs (latest per model × harness — solved sorted first)

ModelHarnessResultTurnstok/sLatencyCtxWhen
gpt-oss-120bbaselinePASS164k2026-06-23
llama-3.3-70bbaselinePASS164k2026-06-23
devstral-small-2-2512baselinePASS164k2026-06-23
mistral-small-3.1-24b-instruct-2503baselinePASS164k2026-06-23
qwen3.6-35b-a3bbaselinePASS64k2026-06-29
qwen3.5-122b-a10b-clusterbaselinePASS64k2026-06-29
qwen3-235b-a22b-q4-clusterbaselinePASS64k2026-06-30
granite-4.1-8bbaselinePASS64k2026-06-30
qwen3.5-4bbaselinePASS64k2026-06-30
deepseek-r1-distill-qwen-32bbaselinePASS64k2026-06-30
granite-4.1-30bbaselinePASS64k2026-06-30
gpt-oss-20bbaselinePASS64k2026-07-01
qwen3-coder-30bbaselinePASS64k2026-07-01
gemma-4-26b-a4b-itbaselinePASS64k2026-07-01
qwen3.6-35b-a3b-orbaselinePASS64k2026-07-01
qwen3.6-35b-bf16baselinePASS64k2026-07-01
gemini-2.5-flash-litebaselinePASS64k2026-07-01
qwen3.6-27bbaselinePASS64k2026-07-01
qwen3.6-27bv1.7.1PASS64k2026-07-02
qwen3-next-80bbaselinePASS64k2026-07-02
phi-4-reasoning-plusbaselinePASS64k2026-07-02
claude-sonnet-4.6baselinePASS64k2026-07-06
qwen3.5-2bbaselinefail64k2026-06-30
qwen3.5-0.8bbaselinefail64k2026-07-01