THE LEADERBOARD

Numbers on
the board.

Overall and category scores are grounded in current, comparable records. Seven-day change needs daily snapshots.

#	Model	Quantum Score	Coding	Reasoning	Research	Agents	Speed	Value
01	Claude Sonnet 5.5Anthropic	88	100	100	—	80	—	65
02	Claude Opus 5.5Anthropic	86	87	97	—	100	—	52
03	GPT-6 AstraOpenAI	60	63	98	—	20	—	22
04	GPT-5.6 SolOpenAI	53	71	69	—	20	—	30
05	Grok 4.6xAI	52	36	65	—	60	—	44
06	Gemini 3.8 FlashGoogle	51	16	62	—	60	—	51
07	GPT-6 SolOpenAI	48	39	67	—	20	—	37
08	Grok 4.5xAI	44	0	90	—	32	—	39
09	GPT-6 LunaOpenAI	Provisional	49	52	—	20	—	48
10	GPT-6.1 SolOpenAI	Provisional	—	—	—	—	—	—
11	GPT-5.6 TerraOpenAI	Provisional	46	67	—	40	—	40
12	GPT-5.6 LunaOpenAI	Provisional	70	51	—	20	—	50
13	Claude Fable 5.1Anthropic	Provisional	43	92	—	40	—	23
14	Claude Haiku 4.5Anthropic	Provisional	—	—	—	—	—	—
15	Gemini 3.7 FlashGoogle	Provisional	44	91	—	20	—	51
16	Gemini 3.6 FlashGoogle	Provisional	42	79	—	20	—	47
17	Gemini 3.5 FlashGoogle	Provisional	41	75	—	20	—	41
18	Gemini 3.5 Flash-LiteGoogle	Provisional	34	0	—	0	—	8
19	Gemini 3.1 Pro PreviewGoogle	Provisional	37	81	—	0	—	34
20	DeepSeek V4.1 FlashDeepSeek	Provisional	55	92	—	100	—	75
21	DeepSeek V4 ProDeepSeek	Provisional	40	95	—	40	—	53
22	Mistral Medium 3.5Mistral	Provisional	—	—	—	—	—	—
23	Mistral Small 4Mistral	Provisional	—	—	—	—	—	—
24	Mistral Large 3Mistral	Provisional	—	—	—	—	—	—
25	Llama 4 ScoutMeta	Provisional	—	—	—	—	—	—
26	Llama 4 MaverickMeta	Provisional	—	—	—	—	—	—
27	Qwen3.5-397B-A17BQwen / Alibaba	Provisional	—	—	—	—	—	—
28	Qwen3.5-122B-A10BQwen / Alibaba	Provisional	—	—	—	—	—	—

Numbers onthe board.