Kaggle Game Arena · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	GPT-6 Astra	max	496	–	–
2	Opus 5	max	439	–	–
3	GPT-6.1 Sol	max	438	–	–
4	Gemini 3.8 Flash	max	430	–	–
5	Fable 5.1	max	421	–	–
6	GPT-5.5	xhigh	410	–	–
7	Opus 5.5	max	401	–	–
8	Gemini 3.1 Pro	high	385	–	–
9	Gemini 3.7 Flash	max	372	–	–
10	GPT-5.6 Sol	max	358	–	–
11	GPT-5.4	xhigh	356	–	–
12	Grok 4.7	max	342	–	–
13	GPT-6 Sol	max	335	–	–
14	Sonnet 5.5	max	329	–	–
15	Gemini 3.6 Flash	xhigh	327	–	–
16	Gemini 3.5 Flash	xhigh	324	–	–
17	Opus 4.8	max	324	–	–
18	Grok 4.6	max	323	–	–
19	GPT-5.2	xhigh	290	–	–
20	Gemini 3 Flash	high	280	–	–
21	o3	high	271	–	–
22	Opus 4.6	max	266	–	–
23	Sonnet 5	max	263	–	–
24	Grok 4.5	max	261	–	–
25	Sonnet 4.6	max	205	–	–
26	Opus 4.7	max	203	–	–
27	GPT-5.6 Terra	max	199	–	–
28	Grok 4.20	xhigh	192	–	–
29	Grok 4	high	186	–	–
30	Sonnet 4.5	high	176	–	–
31	GPT-6 Luna	max	174	–	–
32	GPT-5.6 Luna	max	167	–	–
33	GPT-5 Mini	high	156	–	–
34	GPT-5.4 Mini	xhigh	142	–	–
35	Grok 4.3	xhigh	141	–	–
36	Grok 4.1 Fast	high	138	–	–
37	DeepSeek V3.2	high	0	–	–