niklaslenzGLM-5.3-Flash in the Gauntlet
2 DGX Sparks 256 GB · TensorFold Python engine v0.6.0 + 107 patches (Mia's pin image), Mia's GLM-5.3-Flash recipe @3fd60dfc (EXL3 4 bpw, DFlash2 drafts, tensor parallel 2)
What happened on the course
Each range is one task with one answer and no retry. Every wrong or missing target is a miss and costs a penalty loop. The race time is thinking, running, shooting and loops together.
| # | Task | Targets | Hits | Thinking | Buzzer | Penalty loops | Range time |
|---|---|---|---|---|---|---|---|
| 1 | River ferry (puzzle, level 1) | ✕✕✕✕✕ | 0 / 5 | 6.3 s | no | 5 (60.0 s) | 73.5 s |
| 2 | Painter (code, level 2) | ✕●●✕✕ | 2 / 5 | 5.0 s | no | 3 (36.0 s) | 48.2 s |
| 3 | Bench seating (puzzle, level 3) | ✕✕✕✕● | 1 / 5 | 4.5 s | no | 4 (60.0 s) | 71.7 s |
| 4 | Snake (code, level 4) | ✕✕✕✕✕ | 0 / 5 | 25.4 s | no | 5 (75.0 s) | 107.6 s |
● hit, ✕ miss. The range time runs from arriving at the mat to leaving the range, loops included. A buzzer means the thinking cap ran out and the game forced the answer.
Every number has its file
Result (result.json) Race log (log.jsonl) Model answers (io.jsonl) System prompt Exact command Verify output Model config
Version: prompt version gauntlet-4, tasks bia-2, rules 1, prompt fingerprint 07fff686, tasks fingerprint c2c0f4f6, forced-answer style notes. Verified: the replay of the race log reproduces the end state and every code answer re-runs to the same verdict. Forced-answer style: how the game asks for the answer when the buzzer sounds (a prefilled answer start, or notes when the engine refuses that). Command and config are normalized: internal addresses appear as 127.0.0.1.
- The recipe, image, mode and settings are the same as on the page. Only the start preparation differs: the file cache was not emptied before the start (no administrator rights on this run), and the start check required 111 GiB of free memory. This affects only the start, not how the server runs.
Who made what
- Recipe and EXL3 quantMia's AI LabX @MiaAI_labGitHubHugging FaceWebsite
- ModelZ.ai (GLM-5.3-Flash)Hugging FaceX @Zai_orgGitHubWebsite
- Draft model (DFlash2)Inco AIX @inco_aiHugging FaceGitHubWebsite
- Engine (TensorFold)Ash HartX @ashxhartGitHubHugging FaceWebsite
- EXL3 formatturboderpX @turboderp_Hugging FaceGitHub
Full v2 results of the matching setup (same model, engine, quant and hardware; the engine build of the race is not recorded in the race files) · How the game test works · All races