Season 1 report · Companion
How the Agents Took Notes
Between sessions, each agent remembered only what it wrote to its own files. This is how all 27 runs did that, from file layout and style to the rules they kept and the ones they lost.
Every notes write and read in the Season 1 transcripts. With 2–3 runs per model, these are descriptions, not tests.
§1What I found
- Sessions
- 2,172
- Notes writes
- 1,563
- Capped sessions that saved nothing
- 92%
- Largest note
- 5.9 KB
- The top two took notes in opposite ways. Sol wrote in 96–98% of its sessions across 13 to 70 files. Argon kept one file and wrote in only 27–45% of its sessions. They finished first and second.
- The 50-call cap is where memory dies. Agents save near the end of a session, so a session cut off by the cap usually saves nothing. No run learned to save earlier.
- A rule only works if it survives the next rewrite. There’s no append, so every rewrite has to copy old rules forward. Rules that got copied got obeyed.
- Agents mistook the harness’s log for their diary. Seven runs from four models tried to write the read-only journal folder, 40 times in all.
- No note-taking habit lines up with results. Rank correlations with quests stay between −0.21 and +0.30.
§2Cut off before it could save
Each session allows 50 tool calls. Agents tend to write their notes as one of their last calls, so a session that runs out of calls first loses what it learned that session. Earlier notes and on-chain progress stay intact.
§3Every run’s notebook
One row per run, models by average quests. Writes is the share of sessions that saved at least one note. Size is the main file each time it was rewritten, first session to last.
| Run | Quests | Main file | Files | Writes | Longest gap | Size over the week | Median |
|---|---|---|---|---|---|---|---|
| gpt-6.1-sol-1 | 43 | progress.md | 25 | 96% | 1 | 2.7k | |
| gpt-6.1-sol-2 | 47 | progress.md | 70 | 98% | 1 | 2.3k | |
| gpt-6.1-sol-3 | 46 | plan.md | 13 | 96% | 2 | 2.0k | |
| gemini-4-argon-1 | 48 | notes.md | 1 | 45% | 11 | 3.0k | |
| gemini-4-argon-2 | 39 | notes.md | 1 | 39% | 32 | 2.4k | |
| gemini-4-argon-3 | 44 | plan.md | 1 | 27% | 51 | 2.7k | |
| opus-5.5-1 | 39 | notes.md | 1 | 83% | 8 | 1.8k | |
| opus-5.5-2 | 35 | notes.md | 1 | 94% | 3 | 1.7k | |
| opus-5.5-3 | 45 | notes.md | 1 | 95% | 2 | 2.5k | |
| deepseek-v4-pro-1 | 28 | state.md | 1 | 24% | 62 | 2.6k | |
| deepseek-v4-pro-2 | 30 | notes.md | 1 | 33% | 33 | 2.3k | |
| deepseek-v4-pro-3 | 27 | PLAN.md | 1 | 63% | 14 | 1.9k | |
| muse-wonder-1 | 22 | state.md | 2 | 36% | 39 | 3.0k | |
| muse-wonder-2 | 27 | notes.md | 6 | 51% | 12 | 2.3k | |
| muse-wonder-3 | 23 | plan.md | 11 | 54% | 13 | 3.4k | |
| sonnet-5.5-1 | 30 | notes.md | 1 | 82% | 2 | 0.9k | |
| sonnet-5.5-2 | 18 | notes.md | 1 | 98% | 1 | 1.1k | |
| sonnet-5.5-3 | 14 | notes.md | 1 | 95% | 1 | 1.0k | |
| gpt-6-luna-1 | 19 | plan.md | 1 | 14% | 74 | 2.0k | |
| gpt-6-luna-2 | 22 | plan.md | 2 | 31% | 16 | 2.2k | |
| grok-4.7-1 | 18 | notes.md | 2 | 64% | 4 | 2.8k | |
| grok-4.7-2 | 14 | state.md | 2 | 79% | 4 | 3.1k | |
| grok-4.7-3 | 17 | state.json | 1 | 71% | 3 | 3.4k | |
| gemini-3.6-flash-1 | 13 | state.json | 53 | 88% | 4 | 0.7k | |
| gemini-3.6-flash-2 | 10 | state.md | 4 | 76% | 2 | 1.6k | |
| haiku-5.5-1 | 10 | state.md | 2 | 100% | 0 | 0.9k | |
| haiku-5.5-2 | 6 | state.md | 1 | 100% | 0 | 0.7k |
Longest gap is the longest run of consecutive sessions with no notes write.
§4Rules that stuck, and rules that slipped
Rules agents wrote for themselves, how many later rewrites still carried them, and how often the agent did the forbidden thing before and after.
| Rule | Run | Written | Rewrites keeping it | Before | After |
|---|---|---|---|---|---|
| ban lens_quests | sonnet-5.5-1 | s2 | 57 of 57 | 1 | 10 |
| ban lens_quests | sonnet-5.5-2 | s20 | 26 of 27 | 13 | 7 |
| ban lens_quests | sonnet-5.5-3 | s7 | 14 of 14 | 5 | 1 |
| ban lens_quests | opus-5.5-2 | s32 | 38 of 38 | 37 | 2 |
| ban lens_quests | opus-5.5-3 | s62 | 63 of 63 | 48 | 4 |
| ban lens_quests | haiku-5.5-1 | s2 | 6 of 104 | 3 | 53 |
| orderbook broken / don't retry | gpt-6.1-sol-1 | s5 | 40 of 66 | 1 | 2 |
| orderbook broken / don't retry | gpt-6.1-sol-2 | s5 | 14 of 145 | 1 | 4 |
| orderbook broken / don't retry | gpt-6.1-sol-3 | s8 | 46 of 63 | 1 | 0 |
| orderbook broken / don't retry | grok-4.7-1 | s32 | 20 of 23 | 25 | 0 |
| orderbook broken / don't retry | grok-4.7-2 | s19 | 13 of 14 | 8 | 4 |
| orderbook broken / don't retry | muse-wonder-1 | s21 | 1 of 22 | 15 | 92 |
| orderbook broken / don't retry | muse-wonder-2 | s8 | 2 of 54 | 1 | 42 |
| orderbook broken / don't retry | muse-wonder-3 | s23 | 5 of 31 | 14 | 70 |
| orderbook broken / don't retry | opus-5.5-2 | s36 | 34 of 34 | 2 | 1 |
| orderbook broken / don't retry | sonnet-5.5-2 | s39 | 4 of 4 | 7 | 0 |
| orderbook broken / don't retry | haiku-5.5-1 | s11 | 3 of 88 | 2 | 7 |
| feed blocked by cooldown | opus-5.5-1 | s38 | 33 of 33 | 9 | 5 |
| feed blocked by cooldown | opus-5.5-2 | s19 | 34 of 52 | 3 | 4 |
| feed blocked by cooldown | opus-5.5-3 | s3 | 54 of 122 | 1 | 9 |
| feed blocked by cooldown | gpt-6.1-sol-2 | s21 | 22 of 128 | 4 | 2 |
| do not level above 15 | gpt-6.1-sol-1 | s32 | 19 of 41 | 1 | 0 |
| do not level above 15 | gpt-6.1-sol-2 | s12 | 33 of 139 | 2 | 2 |
| do not level above 15 | gpt-6.1-sol-3 | s53 | 9 of 19 | 2 | 0 |
| 3997 does not exist | gpt-6.1-sol-2 | s12 | 8 of 139 | 2 | 0 |
| 3997 does not exist | gpt-6.1-sol-3 | s21 | 5 of 51 | 3 | 0 |
| 3997 does not exist | opus-5.5-3 | s7 | 0 of 118 | 2 | 5 |
Before and after count the calls the rule was about; for the level rule, level-ups past 15.
§5Model by model
GPT-6.1 Sol 43 · 47 · 46 quests
Wrote in almost every session and spread memory over 13 to 70 files, with a hub file pointing at one file per session. Squashed shorthand: only 5–6% of its characters are spaces, and it almost never says “I” or “we” (one “We” all week).
“Lastsession failed to update memory, hence session5 had to rediscover.”
gpt-6.1-sol-1 · session 5 · the only note in S1 that names its own memory loss (s4 hit the 50-call cap)
“3997doesNOTexist.”
gpt-6.1-sol-2 · session 12 · true, and obeyed (0 later tries), but the free revive quest SQ997 was really index 10003, which this run never found
Gemini 4 Argon 48 · 39 · 44 quests
One notes.md per run, long markdown with bold and headings, but skipped writing in more than half its sessions. Argon #3 wrote its first note at session 52.
“Note: Pool swaps do NOT increment ITEM_TOTAL!”
gemini-4-argon-1 · session 9 · first notes write came at s9; the shouted mechanic note is typical Argon
“Combined spot rate on Node 26: `+136.62/hr`!”
gemini-4-argon-2 · session 121 · dense per-kami rate arithmetic with exclamation marks
Claude Opus 5.5 39 · 35 · 45 quests
One notes.md per run. Reached its working size in the first quarter of the week, then pruned to hold it. Lessons appear right after losses.
“Feed blocked during cooldown (~3 min after start).”
opus-5.5-1 · session 38 · rule kept in all 33 later rewrites, yet 5 more cooldown-blocked feeds followed
“DO MOVES SEQUENTIALLY with wait 4s between — parallel calls fail validation”
opus-5.5-2 · session 72 · operational rule in caps inside a quest bullet
DeepSeek V4 Pro 28 · 30 · 27 quests
state.md, notes.md or PLAN.md, with headed sections. DeepSeek #1 went 62 sessions without a write, the longest silence after Luna #1’s 74.
“## ROOT CAUSE OF STUCK (30 sessions): both kamis dead => no harvest => no MUSU”
deepseek-v4-pro-1 · session 86 · first write after a 62-session silence (s24-85) opens with a post-mortem
“Lesson: 12h > 9.4h max HP, MUST feed mid-run.”
deepseek-v4-pro-2 · session 93 · lesson written the session the kami died mid-quest
Muse Wonder 22 · 27 · 23 quests
Thinks inside its notes, writing “Wait:”, “??” and “hmm” about ten times more than any other model, and says almost nothing aloud.
“## CRITICAL GAS TRAP (unchanged s85->s86)”
muse-wonder-1 · session 86 · alarm heading for a soft-lock
“So counted Scrap rooms = two of {34,26,58,65?,47} but 47 not counted?? Wait:”
muse-wonder-2 · session 80 · thinking out loud inside the notes file
Claude Sonnet 5.5 30 · 18 · 14 quests
One small notes.md, kept tight. Sonnet #1 carried its rule against the full quest list through all 57 rewrites.
“NEVER call lens_quests (huge, 80KB)”
sonnet-5.5-1 · session 62 · the top line of every rewrite; 57/57 later writes kept it
“NEVER call lens_quests (huge, 65KB+ even without full).”
sonnet-5.5-2 · session 23 · same ban as Sonnet-1, found independently
GPT-6 Luna 19 · 22 quests
plan.md, written late. Luna #1 saved its first plan at session 75, about 4.4 days in.
“Session 75 (2026-10-23): Account 3438 KamiSeeker, room 5, 100 stamina, 47 MUSU. Only kami 9714 is DEAD”
gpt-6-luna-1 · session 75 · its first note, 74 sessions in, written as one run-on paragraph
“Do not send doomed harvest/scavenge.”
gpt-6-luna-2 · session 97 · a rule buried mid-paragraph in prose notes
Grok 4.7 18 · 14 · 17 quests
Rulebooks: two to three times the rule words of any other model, in notes.md, state.md or a state.json.
“always dry-run travel_to_room before assuming adjacency.”
grok-4.7-1 · session 58 · rule distilled from a failed route
“Do not buy a kami. Do not bridge. Do not retry get_item_orderbook (bootstrap cache missing)”
grok-4.7-2 · session 33 · three prohibitions in one identity line
Gemini 3.6 Flash 13 · 10 quests
Tried to write the harness’s read-only journal, then Flash #1 copied it into 52 files of its own and read them back 88 times. Flash #1 also kept the smallest notes of the season.
“"status": "bridging_eth_abandoned_monitoring"”
gemini-3.6-flash-1 · session 123 · JSON memory reduced to a status string; its notes list the kami as DEAD from s64 on
“## Mina GDA Price Monitoring (Red Ribbon Gummy - Item 11001)”
gemini-3.6-flash-2 · session 38 · late notes became a price log while the kami stayed dead
Claude Haiku 5.5 10 · 6 quests
Tiny state.md files and a habit of trying the read-only journal folder. The only model that never hit the call cap, so it never lost a session’s notes.
“Don't re-read lens_quests unfiltered (huge output); use quest_state(2).”
haiku-5.5-1 · session 2 · written in s2, kept in only 6 of 104 later writes; 53 more lens_quests calls followed, each ~65 KB
“Session 32-37: blocked. Needs user OK to spend real ETH (bridge/KamiSwap) or other MUSU source.”
haiku-5.5-2 · session 38 · notes waiting for a human who never reads them
Back to the Season 1 report.