Results / LFM2.5-24B-A2B Q4_K_M (llama.cpp, expert offload)

LFM2.5-24B-A2B Q4_K_M (llama.cpp, expert offload)

hermes-lfm25llamacppomarchy (RTX 3060 12 GB)steered

The original prompt had no render check, so the first pass of this build was never checked for things like see-through blocks. We added the check (checklist 1.1.0) after the run, re-checked the top builds, and resumed this same session with plain-words requests about what the check found (see Steers below). Wall time covers both parts; the other numbers on this page are for the resume.

Screenshot of the build

▶ Launch

The game opens in a new tab. It runs sandboxed, so a game that saves to browser storage may not keep its save.

Metrics

Checklist score
22.2 / 100
Design (rubric, 0–5)
–
Stop reason
crash
Model time
10hr 35mins
Tool time (wall minus model)
15mins
Wall time
10hr 51mins
Tool calls
299
Tool errors
125
Errors by type
{"other": 113, "edit_mismatch": 5, "not_found": 1, "timeout": 1, "syntax": 5}
Loop-guard events
0
Tokens in / out
10,177,994.0 / 306,390.0
Thinking tokens
–
Decode / prefill tok/s
23.1 / 613.9
Prefix reuse
0.938
Peak VRAM (GPU / engine)
10,900 / 10,892 MiB
Peak engine RAM
19,856 MiB
Engine
llamacpp
Model
LFM2.5-24B-A2B Q4_K_M (llama.cpp, expert offload) (lfm25-24b-a2b)
Settings
{"thinking": "always on (LFM2.5 template opens <think>), server --reasoning-budget 3072", "reasoning_budget_tokens": 3072, "max_tokens": 7168, "sampling": "LFM2.5 card: temp 0.1, top_k 50, repeat_penalty 1.05 (no top_p/min_p)", "context": "65536 (q8_0 KV, --fit on expert offload), 32768 fallback", "tool_format": "LFM2 pythonic, parsed by llama.cpp (--jinja)", "research": "research/lfm2.5-24b-a2b.md", "max_tokens_boost_ceiling": 16384, "max_tokens_route": "agent cap (hermes_learn --max-tokens), not extra_body (2026-10-09)"}
Node
omarchy (RTX 3060 12 GB)
Prompt
creative-v4
Started
2026-10-17T05:05:16Z

Feature scores

O
0.5
C1
0.5
C4
0
C2
0
C3
0
C5
0

Screenshots

Landing page at desktop and phone widths, and the game at noon, at night, after breaking a block, and on a phone.

landing-desktop.png
landing desktop
landing-hero.png
landing hero
landing-mobile.png
landing mobile

Checklist (3 pass, 2 partial, 13 fail)

FeatureTestScoreEvidence
Ofiles_present1{"required": ["index.html", "game.html"], "notes_md": true}
C1landing_loads1{"status": 200, "errors": [], "network_failed": []}
C1hero_preview0.5{"std": 6.6, "frame_diff": 0, "size": [1280, 720]}
C1gallery0{"images": 3, "generated_nonblank": 0, "distinct": 3, "detail": []}
C1landing_content1{"h1": "Veil Run", "features": 5, "controls": 133, "play": false}
C1play_now0{"error": "no a#play-now"}
C1responsive_mobile0.5{"no_horizontal_scroll": false, "play_now_fits": true, "title_fits": true, "inner": 390, "scroll": 420, "play": [0, 135, 135], "h1": [100, 290, 189]}
C4game_loads0{"exception": "Waiting failed: 25000ms exceeded"}
C2canvas_renders0{"skipped": "game did not load"}
C2fps0{"skipped": "game did not load"}
C2input_responds0{"skipped": "game did not load"}
C2game_states0{"exception": "Waiting failed: 25000ms exceeded"}
C2deterministic0{"exception": "Waiting failed: 25000ms exceeded"}
C3stays_smooth0{"exception": "Waiting failed: 25000ms exceeded"}
C2touch0{"exception": "Waiting failed: 25000ms exceeded"}
C5ground_contact0{"exception": "Waiting failed: 25000ms exceeded"}
C5render_stable0{"exception": "Waiting failed: 25000ms exceeded"}
Oself_contained0{"no_external_requests": true, "pinned_three": false, "no_third_party_branding": true, "external": [], "branded_files": []}

Suite c1.1.2 · Chrome/152.0.7977.82 · WebGL: ANGLE (Intel, Vulkan 1.4.354 (Intel(R) UHD Graphics 630 (CML GT2) (0x00009BC8)), Intel open-source Mesa driver)

Steers (1)

  1. #None · · start · score before

    There is a Play button now, thank you, but the game page still never starts and the front page shows an error. Could you build the game part again as a much simpler version, one small scene where I can move around, and check it really opens before adding more? The page also still slides sideways on my phone.