Name: DuelLab GameBench 2 public release gb2-20260824-e05a1ff89a26
Creator: DuelLab
License: https://creativecommons.org/licenses/by/4.0/

Game profile

Tumbleweed

Tumbleweed is a hex-grid territory game where players grow visible stacks and final control is determined by line-of-sight strength.

Scores here compare player programs using DuelLab's encoded rules for Tumbleweed; they are not a broad measure of human play strength.

By model and setting

GB2-08: Tumbleweed by model and setting

Player programs are ranked by relative per-game score (0–100). Raw rating is shown as an advanced metric, alongside wins, draws, losses, and a per-game uncertainty index.

Reasoning level: Baseline Game: GB2-08: Tumbleweed

GB2-08: Tumbleweed - Baseline filter
Rank	Model	Score	Contributing programs	Raw Elo	W / D / L	Uncertainty
1	Kimi K3	98.9	1	2018.0	165 / 0 / 7	0.0
2	Grok 4.5	85.5	1	1890.8	142 / 0 / 26	0.0
3	GPT-5.6 Sol	75.1	1	1801.6	69 / 10 / 13	8.1
4	Claude Opus 4.8	73.8	1	1792.9	58 / 0 / 20	12.5
5	Claude Opus 4.5	70.1	1	1757.3	57 / 3 / 18	12.5
6	Inkling	69.0	1	1741.3	61 / 7 / 32	6.0
7	Claude Opus 5	67.7	2	1724.8	190 / 14 / 54	3.0
8	Muse Spark 1.2	59.4	1	1652.4	53 / 4 / 33	8.7
9	GLM-5.2	50.2	2	1571.8	45 / 54 / 29	18.5
10	DeepSeek V4 Pro	45.2	1	1515.1	8 / 73 / 15	7.0
11	Minimax M3	44.0	1	1506.5	36 / 5 / 43	10.5
12	GPT-5.4 Nano	43.1	2	1497.1	43 / 86 / 43	10.2
13	Claude Sonnet 5	41.5	1	1482.1	25 / 34 / 27	9.9
14	Qwen3.8 Max	41.5	1	1535.5	3 / 13 / 0	77.5
15	GPT-5.5	36.3	1	1429.4	3 / 76 / 17	7.0
16	Hy3 Preview	36.0	1	1425.0	3 / 83 / 20	4.6
17	Nemotron 3 Ultra 550B A55B	35.8	1	1430.8	1 / 68 / 3	14.8
18	Gemma 4 31B	35.5	1	1419.9	3 / 82 / 21	4.6
19	DeepSeek V4 Flash	35.0	1	1416.8	4 / 80 / 16	6.0
20	Nex N2 Pro	34.2	1	1409.5	0 / 87 / 11	6.5
21	GPT 4.1	34.1	1	1407.8	11 / 72 / 15	6.5
22	North Mini Code	33.9	1	1403.7	4 / 86 / 20	3.7
23	GPT-5.4 Mini	33.5	1	1407.8	0 / 66 / 12	12.5
24	LongCat 2.0	32.3	2	1389.5	6 / 181 / 21	5.0
25	GPT-5.6 Terra	32.2	1	1391.9	2 / 64 / 22	9.2
26	Inkling Small	32.2	1	1452.7	0 / 10 / 4	85.3
27	Laguna S 2.1	32.1	2	1387.6	6 / 179 / 29	4.4
28	Ling 3.0 Flash	31.3	2	1374.3	5 / 244 / 35	0.0
29	Qwen3.7 Max	31.1	1	1379.5	2 / 66 / 30	6.5
30	Gemini 3.6 Flash	30.5	2	1364.2	16 / 226 / 92	0.0
31	Mistral Medium 3.5	30.1	1	1368.3	1 / 87 / 18	4.6
32	GPT-5.4	30.0	1	1369.5	5 / 55 / 35	7.3
33	Gemini 3.5 Flash Lite	29.0	2	1350.4	14 / 225 / 92	0.0
34	O3	28.5	1	1346.5	5 / 105 / 47	0.0
35	MiMo-V2.5	28.0	2	1353.8	2 / 120 / 44	10.8
36	Ox Alpha	26.4	2	1398.1	0 / 11 / 17	85.3
37	Step 3.7 Flash	22.8	1	1292.5	0 / 99 / 51	0.0
38	GPT-OSS 120B	20.5	1	1270.6	3 / 89 / 58	0.0
39	GPT-5.6 Luna	0.0	1	1083.9	4 / 12 / 74	8.7

Underlying results

Individual player programs

Each row is one accepted player program for GB2-08: Tumbleweed. A model can appear more than once when it produced more than one program. Score is the normalized 0–100 result; Elo and W / D / L show the underlying play record. Confidence and evidence can be blank when too few compatible matches are available. Draws at the move limit count; one-sided model faults are listed separately.

Search model Status Sort Showing 51 player programs.

Player ID	Model	Status	Score	Elo	Confidence	W / D / L	Rated / attempted	Opponents	Capped draws	Faults / players	Generation
E-01b57d5e6655f927	Claude Opus 4.5	Rated	70.1	1757	—	57 / 3 / 18	78 / 78	39	3	0 / 1	First try · 2026-07-02
E-c9d7578e8375d994	Claude Opus 4.8	Rated	73.8	1793	—	58 / 0 / 20	78 / 78	39	0	0 / 1	First try · 2026-07-02
E-43cc1c89fa5720b3	Claude Opus 5	Rated	79.9	1838	—	124 / 14 / 20	158 / 158	60	14	0 / 1	First try · 2026-07-25
E-5a7942274854aaa6	Claude Opus 5	Rated	55.4	1611	—	66 / 0 / 34	100 / 100	50	0	0 / 1	First try · 2026-07-25
E-c2e7d7e507756f84	Claude Sonnet 5	Rated	41.5	1482	—	25 / 34 / 27	86 / 86	42	34	0 / 1	First try · 2026-08-09
E-c36a98e057c264eb	North Mini Code	Rated	33.9	1404	—	4 / 86 / 20	110 / 110	48	86	0 / 1	First try · 2026-07-11
E-4375cf96eeb624e2	DeepSeek V4 Flash	Rated	35.0	1417	—	4 / 80 / 16	100 / 100	48	80	0 / 1	Repaired · 2026-07-02
E-c296a7abab6975d7	DeepSeek V4 Pro	Rated	45.2	1515	—	8 / 73 / 15	96 / 96	47	73	0 / 1	First try · 2026-07-02
E-5aa593158b0d8b3a	Gemini 3.5 Flash	Waiting	—	—	—	0 / 0 / 0	0 / 0	0	0	0 / 1	First try · 2026-07-21
E-284d554eca16e6f6	Gemini 3.5 Flash Lite	Rated	29.5	1355	—	7 / 112 / 47	166 / 166	61	112	0 / 1	First try · 2026-07-24
E-af119b49b8f1d63d	Gemini 3.5 Flash Lite	Rated	28.6	1346	—	7 / 113 / 45	165 / 165	60	113	0 / 1	First try · 2026-07-24
E-657c523bc9e9f80c	Gemini 3.6 Flash	Rated	31.0	1369	—	7 / 111 / 48	166 / 166	61	111	0 / 1	Repaired · 2026-07-24
E-c694a36a5e4e3f09	Gemini 3.6 Flash	Rated	30.0	1359	—	9 / 115 / 44	168 / 168	62	115	0 / 1	First try · 2026-07-24
E-8546650b7690a81a	Gemma 4 31B	Rated	35.5	1420	—	3 / 82 / 21	106 / 106	51	82	0 / 1	First try · 2026-07-02
E-eabc996d553eebe0	GPT 4.1	Rated	34.1	1408	—	11 / 72 / 15	98 / 98	46	72	0 / 1	First try · 2026-07-02
E-ae3a897b3420bcfb	GPT 5	Waiting	—	—	—	0 / 0 / 0	0 / 0	0	0	0 / 1	First try · 2026-07-21
E-07ad6c481bf678a2	GPT-5.4	Rated	30.0	1370	—	5 / 55 / 35	95 / 95	42	55	0 / 1	Repaired · 2026-08-09
E-bb8ac98c284a38ce	GPT-5.4 Mini	Rated	33.5	1408	—	0 / 66 / 12	78 / 80	39	66	2 / 1	First try · 2026-07-02
E-024aff0a983df40d	GPT-5.4 Nano	Rated	31.4	1383	—	0 / 82 / 16	98 / 98	48	82	0 / 1	First try · 2026-07-09
E-fd5a36e5fd795426	GPT-5.4 Nano	Rated	54.7	1611	—	43 / 4 / 27	74 / 74	37	4	0 / 1	Repaired · 2026-07-02
E-c42fbee03a029c83	GPT-5.5	Rated	36.3	1429	—	3 / 76 / 17	96 / 96	46	76	0 / 1	First try · 2026-07-02
E-b761803f2b99649c	GPT-5.6 Luna	Rated	0.0	1084	—	4 / 12 / 74	90 / 90	43	12	0 / 1	First try · 2026-08-09
E-9c0688f3432e6295	GPT-5.6 Sol	Rated	75.1	1802	—	69 / 10 / 13	92 / 92	42	10	0 / 1	First try · 2026-08-09
E-44cace55a6126cdf	GPT-5.6 Terra	Rated	32.2	1392	—	2 / 64 / 22	88 / 90	44	64	2 / 1	First try · 2026-07-10
E-11d1ae7d9526f835	Ling 3.0 Flash	Rated	31.9	1381	—	0 / 126 / 14	140 / 144	52	126	0 / 1	First try · 2026-07-24
E-28bcc8394e411616	Ling 3.0 Flash	Rated	30.6	1368	—	5 / 118 / 21	144 / 144	55	118	0 / 1	Repaired · 2026-07-24
E-87d69837bf65e78b	LongCat 2.0	Rated	30.9	1376	—	1 / 93 / 10	104 / 104	49	93	0 / 1	First try · 2026-07-24
E-df8e0fab4bf8bb33	LongCat 2.0	Rated	33.6	1403	—	5 / 88 / 11	104 / 104	49	88	0 / 1	First try · 2026-07-24
E-20b4e41aba8e4f47	Muse Spark 1.2	Rated	59.4	1652	—	53 / 4 / 33	90 / 90	42	4	0 / 1	First try · 2026-08-13
E-e67e41d4491d7ad3	Minimax M3	Rated	44.0	1506	—	36 / 5 / 43	84 / 84	42	5	0 / 1	First try · 2026-07-02
E-ba85be2fa53925f5	Mistral Medium 3.5	Rated	30.1	1368	—	1 / 87 / 18	106 / 106	53	87	0 / 1	First try · 2026-07-02
E-3e1be56b9d966ded	Kimi K3	Rated	98.9	2018	—	165 / 0 / 7	172 / 172	67	0	0 / 1	First try · 2026-07-24
E-506d972b5c69a00d	Nex N2 Pro	Rated	34.2	1410	—	0 / 87 / 11	98 / 98	48	87	0 / 1	Repaired · 2026-07-11
E-ada32b27ea6ffedd	Nemotron 3 Ultra 550B A55B	Rated	35.8	1431	—	1 / 68 / 3	72 / 72	36	68	0 / 1	First try · 2026-07-02
E-335ae90fdc702bd6	O3	Rated	28.5	1347	—	5 / 105 / 47	157 / 157	61	105	0 / 1	First try · 2026-07-21
E-f6e9da46c61ad315	GPT-OSS 120B	Rated	20.5	1271	—	3 / 89 / 58	150 / 150	62	89	0 / 1	First try · 2026-07-21
E-4030acdbccf3ca03	Laguna S 2.1	Rated	31.7	1383	—	3 / 90 / 13	106 / 106	50	90	0 / 1	Repaired · 2026-07-24
E-e90a3ce84847ba43	Laguna S 2.1	Rated	32.6	1392	—	3 / 89 / 16	108 / 108	51	89	0 / 1	First try · 2026-07-24
E-a667eb7d15f7f43e	Qwen3.7 Max	Rated	31.1	1379	—	2 / 66 / 30	98 / 98	48	66	0 / 1	First try · 2026-07-02
E-e77a22943f077d4d	Qwen3.8 Max	Limited data	41.5	1536	—	3 / 13 / 0	16 / 16	7	13	0 / 1	First try · 2026-08-08
E-4e4d3135cae0253d	Ox Alpha	Rated	25.9	1393	—	0 / 5 / 9	14 / 14	7	5	0 / 1	First try · 2026-08-22
E-6b99e270ec28f9cf	Ox Alpha	Rated	27.0	1403	—	0 / 6 / 8	14 / 14	7	6	0 / 1	First try · 2026-08-22
E-c38fe2712e69333b	Step 3.7 Flash	Rated	22.8	1293	—	0 / 99 / 51	150 / 150	62	99	0 / 1	First try · 2026-07-21
E-56487337fad4192f	Hy3 Preview	Rated	36.0	1425	—	3 / 83 / 20	106 / 106	51	83	0 / 1	First try · 2026-07-02
E-68669040dc891a7d	Inkling	Rated	69.0	1741	—	61 / 7 / 32	100 / 100	48	7	0 / 1	Repaired · 2026-07-24
E-99211d414b6fcc51	Inkling Small	Limited data	32.2	1453	—	0 / 10 / 4	14 / 14	6	10	0 / 1	Repaired · 2026-08-09
E-b914d026c7631fcb	Grok 4.5	Rated	85.5	1891	—	142 / 0 / 26	168 / 168	69	0	0 / 1	First try · 2026-07-24
E-cabbf604f15eb575	MiMo-V2.5	Rated	23.7	1312	—	2 / 54 / 30	86 / 86	43	54	0 / 1	First try · 2026-07-02
E-d51fce0298267f96	MiMo-V2.5	Rated	32.4	1396	—	0 / 66 / 14	80 / 80	40	66	0 / 1	First try · 2026-07-01
E-6a361d88d8c1fb48	GLM-5.2	Rated	34.3	1417	—	3 / 54 / 13	70 / 70	35	54	0 / 1	First try · 2026-07-11
E-97efd9455394f6d7	GLM-5.2	Rated	66.1	1726	—	42 / 0 / 16	58 / 58	29	0	0 / 1	Repaired · 2026-07-02