aitest index

generated 2026-10-24 13:16 · IDEA.md · SCORING.md · RANKING.md
runs
57
provider
harness
tier
tiermodelparamsproviderharnessdatetimeout tokenstpsindex.html
S#1▶ Playclaude-fable-5-1
rules 4 feels right · effects 4 changes every 10 lines · sound 4 BGM synced with effects
-anthropicclaude-code2026-09-02 21:2527m53s122.4k8258.5 KB
A#2▶ Playclaude-opus-5-5
rules 4 feels right · effects 4 changes every 10 lines · sound 3 has BGM
-anthropicclaude-code2026-09-23 10:3214m30s88.6k11683.2 KB
A#3▶ Playclaude-opus-5-5
rules 4 feels right · effects 4 changes every 10 lines · sound 3 has BGM
-anthropicclaude-code2026-10-14 10:298m45s58.4k11452.3 KB
A#4▶ Playclaude-sonnet-5-5
rules 4 feels right · effects 4 changes every 10 lines · sound 3 has BGM
-anthropicclaude-code2026-09-29 23:284m25s38.7k15241.6 KB
A#5▶ PlayMiMo-v2.6pro
rules 4 feels right · effects 4 changes every 10 lines · sound 3 has BGM
Result received from someone else (index.html only)
1020B-A42B-opencode2026-09-23 10:4636m14s--49.7 KB
A#6▶ PlayQwen3.8-Flash-Next
Qwen3.8-Flash-Next-GGUF/UD-IQ4_XS/Qwen3.8-Flash-Next-UD-IQ4_XS-00001-of-00003.gguf
rules 4 feels right · effects 4 changes every 10 lines · sound 3 has BGM
index.html is built by concatenating parts/ with build.sh. Also includes test.html and test/ (node tests)
177B-A7Bllama.cpppi-agent2026-09-05 16:586h54m259.8k1187.0 KB
A#7▶ Playclaude-opus-5
rules 4 feels right · effects 3 screen explodes · sound 4 BGM synced with effects
-anthropicclaude-code2026-09-02 22:1815m56s77.1k8462.6 KB
A#8▶ Playgpt-6-astra
rules 4 feels right · effects 3 screen explodes · sound 4 BGM synced with effects
Result received from someone else (index.html only). No session log
-openaicodex2026-09-07 10:5110m19s--34.1 KB
A#9▶ PlayGLM-5.3-Flash (max thinking)
rules 4 feels right · effects 3 screen explodes · sound 3 has BGM
Result received from someone else (index.html only). Provider unknown; harness opencode; GLM-5.3-Flash with max thinking
321B-A17B-opencode2026-09-09 02:3155m38s--56.5 KB
A#10▶ PlayGrok 4.6 (Extra High)
rules 4 feels right · effects 3 screen explodes · sound 3 has BGM
Result received from someone else (index.html only)
--opencode2026-09-13 01:3729m41s--63.3 KB
A#11▶ PlayMiniMax-M3.1-Flash (Space Bunny, Max think)
rules 4 feels right · effects 3 screen explodes · sound 3 has BGM
Received as a stealth model under the codename Space Bunny; later revealed to be MiniMax-M3.1-Flash (https://note.com/humble_bobcat51/n/n0b318ab738ec). Result received from someone else (index.html only)
428B-A23B--2026-09-24 02:5043m21s--55.7 KB
A#12▶ PlayQwen3.8-27B (EXL3 3.5bpw)
Qwen3.8-27B-EXL3-3.5bpw
rules 4 feels right · effects 3 screen explodes · sound 3 has BGM
Result received from someone else (index.html only). Reportedly ~300K tokens in+out, but the harness stats were broken
27Bexllamav3Deepseek Harness2026-09-16 05:351h02m--37.6 KB
A#13▶ Playclaude-sonnet-5
rules 4 feels right · effects 2 clearly visible effects · sound 3 has BGM
-anthropicclaude-code2026-09-02 22:4417m20s87.8k10050.0 KB
A#14▶ PlayDeepSeek-V4.1-Flash (1017)
rules 4 feels right · effects 2 clearly visible effects · sound 2 sound per action, varies on clear
Served by OpenCode's own provider (free tier, model id deepseek-v4.1-flash-free)
552B-A16Bopencodeopencode2026-10-20 14:581h14m190.3k43103.5 KB
B#15▶ Playclaude-haiku-5-5
rules 4 feels right · effects 1 slight glow · sound 3 has BGM
-anthropicclaude-code2026-10-08 10:526m27s60.5k24031.0 KB
B#16▶ PlayOrnith-1.5-122B
Ornith-1.5-122B-GGUF/UD-Q4_K_XL/Ornith-1.5-122B-UD-Q4_K_XL-00001-of-00003.gguf
rules 4 feels right · effects 2 clearly visible effects · sound 1 sound effects but monotonous
122B-A9Bllama.cpppi-agent2026-10-24 12:4210m53s13.3k2027.2 KB
B#17▶ PlayBeam
rules 4 feels right · effects 1 slight glow · sound 3 has BGM
Stealth model of unknown origin served by OpenCode's own provider (codename beam-free). Details to be added once revealed
-opencodeopencode2026-10-20 18:225m34s45.4k13629.3 KB
B#18▶ PlayDeepSeek-V4-Flash
DeepSeek-V4-Flash-0731-GGUF/UD-IQ3_XXS/DeepSeek-V4-Flash-0731-UD-IQ3_XXS-00001-of-00004.gguf
rules 4 feels right · effects 1 slight glow · sound 0 silent
284B-A13Bllama.cpppi-agent2026-09-05 02:0758m34s48.0k1432.6 KB
B#19▶ Playgpt-5.5
rules 4 feels right · effects 3 screen explodes · sound 0 silent
Layout overflows the screen
-openaicodex2026-09-03 22:304m40s13.7k4930.9 KB
B#20▶ PlayGrok 4.7 (High Fast)
rules 3 complete but controls feel off · effects 4 changes every 10 lines · sound 3 has BGM
Result received from someone else (index.html only)
bugs: Key repeat rate is effectively infinite
--opencode2026-09-22 02:3435m45s--75.9 KB
B#21▶ PlayMiMo-V2.6-Flash
rules 3 complete but controls feel off · effects 3 screen explodes · sound 2 sound per action, varies on clear
Served by OpenCode's own provider (free tier, model id mimo-v2.6-flash-free)
bugs: Rotating while holding an arrow key cancels the arrow input
309B-A15Bopencodeopencode2026-10-07 01:1953m56s142.8k4484.6 KB
B#22▶ Playgemma-4-26B-A4B-it
gemma-4-26B-A4B-it-GGUF/gemma-4-26B-A4B-it-UD-Q5_K_M.gguf
rules 4 feels right · effects 1 slight glow · sound 1 sound effects but monotonous
bugs: Moving requires repeated key presses
26B-A4Bllama.cpppi-agent2026-09-02 21:434m52s18.4k6324.1 KB
B#23▶ PlayHy4 preview
rules 2 playable but missing features · effects 3 screen explodes · sound 3 has BGM
Result received from someone else (index.html only)
bugs: Core rules seem right, but the next-piece panel and other displays are buggy. Occasionally misses key events
770B-A49B-opencode2026-09-10 01:1021m14s--61.4 KB
B#24▶ PlayDeepSeek-V4.1-Flash
rules 3 complete but controls feel off · effects 4 changes every 10 lines · sound 2 sound per action, varies on clear
Result received from someone else (index.html only)
bugs: Arrow-key input that starts during the line-clear effect is ignored
Effects are more solid than GLM-5.3-Flash
552B-A16B-opencode2026-09-11 11:36---47.4 KB
C#25▶ Playgpt-oss-120b
gpt-oss-120b-GGUF/UD-Q4_K_XL/gpt-oss-120b-UD-Q4_K_XL-00001-of-00002.gguf
rules 2 playable but missing features · effects 1 slight glow · sound 0 silent
117B-A5Bllama.cpppi-agent2026-09-05 19:480m58s2.5k435.0 KB
C#26▶ PlayQwen3-Next-80B-A3B-Thinking
Qwen3-Next-80B-A3B-Thinking-GGUF/Qwen3-Next-80B-A3B-Thinking-Q4_K_M.gguf
rules 2 playable but missing features · effects 0 no effects · sound 0 silent
80B-A3Bllama.cpppi-agent2026-09-05 23:533m31s10.3k498.1 KB
C#27▶ PlayQwen3.8-27B
Qwen3.8-27B-GGUF/Qwen3.8-27B-UD-IQ4_XS.gguf
rules 1 fatal bug · effects 4 changes every 10 lines · sound 4 BGM synced with effects
bugs: Pieces sometimes deform on rotation
27Bllama.cpppi-agent2026-09-02 22:552h54m116.7k1258.8 KB
C#28▶ PlayMuse Spark 1.3
rules 1 fatal bug · effects 4 changes every 10 lines · sound 3 has BGM
Result received from someone else (index.html only)
bugs: When clearing two lines at once, one of them remains
---2026-09-10 00:522m53s--32.8 KB
C#29▶ Playgemini-3.8-flash-high
rules 1 fatal bug · effects 4 changes every 10 lines · sound 2 sound per action, varies on clear
bugs: Occasionally a piece does not lock after reaching the bottom
-googleantigravity-cli2026-09-08 21:5722m45s--87.9 KB
C#30▶ Playgpt-5.6-sol
rules 1 fatal bug · effects 4 changes every 10 lines · sound 1 sound effects but monotonous
Result received from someone else (index.html only). No session log
bugs: Board sometimes gets corrupted when lines are cleared
-openaicodex2026-09-07 12:139m22s--25.4 KB
C#31▶ Playgemini-3.7-flash-high
rules 1 fatal bug · effects 3 screen explodes · sound 3 has BGM
bugs: Sometimes lines are not removed when clearing multiple lines at once
-googleantigravity-cli2026-09-08 23:467m16s--81.2 KB
C#32▶ Playgemini-3.6-flash-high
rules 1 fatal bug · effects 3 screen explodes · sound 2 sound per action, varies on clear
bugs: Board gets corrupted when lines are cleared
-googleantigravity-cli2026-09-08 23:522m57s--48.4 KB
C#33▶ Playgpt-5.6-terra
rules 1 fatal bug · effects 2 clearly visible effects · sound 1 sound effects but monotonous
bugs: Moving requires repeated key presses. Board breaks when clearing multiple lines at once
-openaicodex2026-09-03 23:302m19s6.3k4513.0 KB
C#34▶ PlayLongCat-2.5-preview
rules 1 fatal bug · effects 1 slight glow · sound 1 sound effects but monotonous
Served by OpenCode's own provider (free tier, model id longcat-2.5-preview-free)
bugs: Clearing two or more lines at once removes them incorrectly
1600B-A48Bopencodeopencode2026-10-07 01:0411m18s29.6k4426.2 KB
C#35▶ PlayFlint v1.0 (40B)
rules 1 fatal bug · effects 1 slight glow · sound 1 sound effects but monotonous
Served by OpenCode's own provider (free tier, model id flint-40b-free)
bugs: Cannot rotate; pieces disappear
40B-A2.5Bopencodeopencode2026-10-22 08:5713m48s76.1k9227.4 KB
C#36▶ PlayFlint v1.0 (40B)
rules 1 fatal bug · effects 1 slight glow · sound 1 sound effects but monotonous
OpenRouter free tier (recursive/flint-v1.0:free)
bugs: Key repeat is immediate and infinitely fast, so pieces cannot be moved
40B-A2.5Bopenrouterpi-agent2026-10-22 09:1111m30s86.7k12625.1 KB
F#37▶ Playgpt-5.6-luna
rules 1 fatal bug · effects 1 slight glow · sound 1 sound effects but monotonous
bugs: Most pieces get corrupted on rotation
-openaicodex2026-09-03 23:001m47s5.0k4611.5 KB
F#38▶ PlayGLM-4.7-Flash
GLM-4.7-Flash-GGUF/GLM-4.7-Flash-UD-Q4_K_XL.gguf
rules 1 fatal bug · effects 1 slight glow · sound 1 sound effects but monotonous
30B-A3Bllama.cpppi-agent2026-09-05 23:292m28s6.5k4425.8 KB
F#39▶ PlayQwen3-Coder-Next
Qwen3-Coder-Next-GGUF/Qwen3-Coder-Next-Q4_K_M.gguf
rules 1 fatal bug · effects NA not assessable · sound 1 sound effects but monotonous
80B-A3Bllama.cpppi-agent2026-09-05 23:444m19s11.3k4449.1 KB
F#40▶ PlayDevstral-2-123B-Instruct
Devstral-2-123B-Instruct-2512-GGUF/Q4_K_M/Devstral-2-123B-Instruct-2512-Q4_K_M-00001-of-00002.gguf
rules 1 fatal bug · effects NA not assessable · sound 0 silent
bugs: Key repeat is too fast
123Bllama.cpppi-agent2026-09-05 19:4825m33s4.3k314.8 KB
F#41▶ Playgpt-5.4-mini
rules 1 fatal bug · effects NA not assessable · sound NA not assessable
bugs: Cannot move pieces
-openaicodex2026-09-03 22:107m06s28.4k6750.1 KB
F#42▶ Playfledge-alpha (max)
rules 0 does not start / unplayable · effects NA not assessable · sound 1 sound effects but monotonous
Stealth model of unknown origin served by OpenCode's own provider (codename fledge-alpha-free, variant max). Details to be added once revealed
-opencodeopencode2026-10-06 22:584m43s15.7k5520.3 KB
F#43▶ PlayGLM-4.5-Air
GLM-4.5-Air-GGUF/GLM-4.5-Air-Q4_K_M-00001-of-00002.gguf
rules 0 does not start / unplayable · effects NA not assessable · sound 0 silent
106B-A12Bllama.cpppi-agent2026-09-05 23:115m16s5.2k1722.0 KB
F#44▶ PlayQwen3.6-35B-A3B
Qwen3.6-35B-A3B-MTP-GGUF/Qwen3.6-35B-A3B-UD-Q4_K_XL.gguf
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
35B-A3Bllama.cpppi-agent2026-09-02 22:334m32s15.0k5539.0 KB
F#45▶ Playclaude-haiku-4-5
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
bugs: Pieces do not stop at the bottom and fall forever
-anthropicclaude-code2026-09-02 23:074m32s22.9k9232.2 KB
F#46▶ Playminimax/minimax-m3:free
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
bugs: Broken HTML: canvas is a div
428B-A23Bopenrouterpi-agent2026-09-03 21:517m05s13.0k10434.0 KB
F#47▶ PlayQwen3-Coder-30B-A3B-Instruct
Qwen_Qwen3-Coder-30B-A3B-Instruct-GGUF/Qwen3-Coder-30B-A3B-Instruct-Q3_K_M.gguf
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
30B-A3Bllama.cpppi-agent2026-09-06 12:281m10s4.8k6919.5 KB
F#48▶ PlayMiniMax-M2.7
MiniMax-M2.7-GGUF/UD-IQ3_S/MiniMax-M2.7-UD-IQ3_S-00001-of-00003.gguf
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
230B-A10Bllama.cpppi-agent2026-09-06 13:128m49s9.1k1730.2 KB
F#49▶ PlayNemotron-3-Super-120B-A12B
NVIDIA-Nemotron-3-Super-120B-A12B-GGUF/UD-IQ4_NL/NVIDIA-Nemotron-3-Super-120B-A12B-UD-IQ4_NL-00001-of-00003.gguf
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
120B-A12Bllama.cpppi-agent2026-09-09 22:179m53s9.0k1513.7 KB
F#50▶ Play▶ fixedDeepSeek-V4.1-Flash
also: fixed.html
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
index.html is the first result. Regulation violation: I then reported a bug and the agent fixed it in 12 more minutes (fixed.html). Stats below include that
552B-A16BdeepseekDeepseek Harness2026-09-16 18:3056m11s293.2k28296.8 KB
F#51▶ PlayMiniMax-M3.1-Flash (Space Bunny)
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
Received as a stealth model under the codename Space Bunny; later revealed to be MiniMax-M3.1-Flash (https://note.com/humble_bobcat51/n/n0b318ab738ec). Result received from someone else (index.html only)
bugs: Cannot rotate
428B-A23B--2026-09-24 00:45---62.6 KB
F#52▶ Playbig-pickle
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
Stealth model of unknown origin served by OpenCode's own provider (codename big-pickle). Details to be added once revealed. The agent made no tool calls and only pasted the HTML in its reply; index.html was restored from that text. The JavaScript is incomplete (ends mid-function with a comment that it is too long to write out) and has a syntax error
bugs: Incomplete JavaScript with a syntax error; does not run
-opencodeopencode2026-10-06 23:0515m36s18.9k2054.3 KB
F#53▶ PlaySarashina3 mini
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
SB Intuitions Sarashina API (api.sarashina.ai). Output is an empty HTML skeleton with placeholder comments instead of code
bugs: Empty HTML skeleton; no game code
-sarashinapi-agent2026-10-14 10:300m12s0.9k730.7 KB
F#54▶ PlaySarashina3 nano
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
SB Intuitions Sarashina API (api.sarashina.ai). Output is an empty HTML skeleton with placeholder comments instead of code
bugs: Empty HTML skeleton; no game code
-sarashinapi-agent2026-10-14 10:310m01s0.3k2430.5 KB
F#55▶ PlayNemotron-3.5-Super-VL
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
OpenRouter free tier (nvidia/nemotron-3.5-super-vl:free)
120B-A12Bopenrouterpi-agent2026-10-18 23:5811m18s8.8k6719.2 KB
F#56▶ PlayNorth Mini Code
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
OpenRouter free tier (cohere/north-mini-code:free)
30B-A3Bopenrouterpi-agent2026-10-19 00:061m02s9.2k14928.7 KB
F#57▶ PlayMiniMax-M3.1-Flash
rules 0 does not start / unplayable · effects NA not assessable · sound NA not assessable
OpenRouter free tier (minimax/minimax-m3.1-flash:free). A unit test the agent ran went into an infinite loop, so the run was killed manually
bugs: Game over the moment it starts
428B-A23Bopenrouterpi-agent2026-10-20 10:2716m48s66.3k15752.0 KB