Champions BSS · own teams (c1) · 70 rated games · updated 2026-10-17
Champions BSS · own teams (c1)
Archived. A superseded version, kept for the record; its games count on no current leaderboard. The current one is Rotom Champ Bench.
The first Champions version (retired after 24 games): the battle prompt didn't say the Pokémon are level 50, and DeepSeek's thinking was cut off by providers that ignore the budget.
Ratings are on the Elo scale, fitted by Bradley-Terry over this board's rated games. Games are only played once; each leaderboard is a different selection of them. Bonsai 2 27B is fixed at 1500 by definition; every other number is relative to it.3Models with fewer than 16 games are marked provisional. Click a column to sort.
| # | Model | Rating (95% CI)1 | Games | W | L | Win rate | Halluc.2 | Tokens / turn | Cost / game |
|---|---|---|---|---|---|---|---|---|---|
| — | Claude Haiku 5.5 | 8 | 3 | 5 | 38% | 1.3% | 926 | $0.009 | |
| — | DeepSeek V4.1 Flash | 8 | 4 | 4 | 50% | 2.2% | 1,109 | $0.012 | |
| — | Gemma 4 31B | 18 | 9 | 9 | 50% | 3.3% | 754 | $0.005 | |
| — | GPT-6 Luna Decisions | 20 | 18 | 2 | 90% | 0.0% | 0 | $0.003 | |
| — | GPT-6 Luna Pro | 10 | 9 | 1 | 90% | 0.0% | 1,380 | $0.016 | |
| — | GPT-OSS 20B | 8 | 5 | 3 | 63% | 3.6% | 1,279 | free | |
| — | Jev 1.13 | 16 | 5 | 11 | 31% | 0.0% | 0 | $0.002 | |
| — | Ling 3.0 Tiny | 10 | 6 | 4 | 60% | 2.8% | 2,352 | free | |
| — | MiMo V2.6 Flash | 8 | 2 | 6 | 25% | 2.4% | 855 | $0.006 | |
| — | MiniCPM5 2B | 10 | 2 | 8 | 20% | 3.4% | 1,607 | free | |
| — | Muse Spark 1.3 | 8 | 4 | 4 | 50% | 0.1% | 1,590 | $0.008 | |
| — | Chimps at the Zoo | 16 | 3 | 13 | 19% | — | — | — |
Rules
Every game records the rules it was played under; a board only rates games that match their version's stored rules.
| Version | Format | Teams | Bring | Thinking per turn | Tokens per turn | Prompt |
|---|---|---|---|---|---|---|
| c1 | gen9championsrotomarenabssregmb | own (drafted) | 3 of 6 | 1,024 tokens | 2,048 | 41db2ebeaa2a · draft b0a500bd511d |
Head to head
Wins–losses of the row model against the column model. Click a record for its games and replays.
| Claude Haiku 5.5 | DeepSeek V4.1 Flash | Gemma 4 31B | GPT-6 Luna Decisions | GPT-6 Luna Pro | GPT-OSS 20B | Jev 1.13 | Ling 3.0 Tiny | MiMo V2.6 Flash | MiniCPM5 2B | Muse Spark 1.3 | Chimps at the Zoo | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Haiku 5.5 | · | · | · | · | · | · | · | |||||
| DeepSeek V4.1 Flash | · | · | · | · | · | · | · | |||||
| Gemma 4 31B | · | · | ||||||||||
| GPT-6 Luna Decisions | · | |||||||||||
| GPT-6 Luna Pro | · | · | · | · | · | · | ||||||
| GPT-OSS 20B | · | · | · | · | · | · | · | |||||
| Jev 1.13 | · | · | · | |||||||||
| Ling 3.0 Tiny | · | · | · | · | · | · | ||||||
| MiMo V2.6 Flash | · | · | · | · | · | · | · | |||||
| MiniCPM5 2B | · | · | · | · | · | · | ||||||
| Muse Spark 1.3 | · | · | · | · | · | · | · | |||||
| Chimps at the Zoo | · | · | · |
- 95% intervals come from 1,000 parametric bootstrap refits. When an interval covers half the chart, the model has not played much yet, or plays very inconsistently, or both.
- Hallucination rate: the share of turns where the model gave no usable answer (named a move it does not have, answered in a format we could not read, rambled past its token budget, timed out or errored) and the default move was played on its behalf. Matches where either side does this on more than 50% of its turns are voided.
- Ratings need a reference point. We chose a local open-weights model because it is always available, costs nothing to run, and cannot be deprecated by its vendor while we sleep.