Quality: the BenchLeader Index

Composite of every quality benchmark, normalised per benchmark and averaged per category. Requires results on at least five benchmarks across two categories.

As of 5 Oct 2026, BenchLeader ranks GPT-6 Astra first on the Quality: the BenchLeader Index board at 70.2, ahead of Claude Opus 5.5 (70.1) and Claude Opus 5.5 (70.1).

756 models · data as of 5 Oct 2026

1. 1GPT-6 Astra (max)70.2
2. 2Claude Opus 5.5 (high)70.1
3. 3Claude Fable 5.1 (xhigh)69.9
4. 4Gemini 4 Argon (high)69.7
5. 5GPT-6.1 Sol (max)69.5
6. 6Claude Fable 5 (thinking)69.2
7. 7Claude Opus 5 (high)68.4
8. 8Muse Spark 1.3 (max)67.9
9. 9GPT-5.6 Sol (max)67.5
10. 10Claude Sonnet 5.5 (xhigh)67.0
11. 11GPT-6 Sol (max)66.3
12. 12GPT-5.5 (xhigh)65.9
13. 13Kimi K3 (max)65.6
14. 14GPT-5.5 Pro (xhigh)65.6
15. 15Qwen3.8 Max (max)65.0

SearchAll providers01.AIAi2AI21 LabsAI9StarsAlibabaAmazonAnt GroupAnthropicApodexArcee AIBaiduByteDanceCelerisChina MobileCohereDeep CogitoDeepSeekGoogleIBMInceptionKTKuaishouKwaiKATLG AI ResearchLiquid AIMBZUAIMeituanMetaMicrosoftMiniMaxMistral AIMoonshot AIMotif TechnologiesMultiverse ComputingNanbeigeNaverNex AGINous ResearchNVIDIAOpenAIOpenBMBOpenChatPoolsidePrime IntellectRekaSarvamServiceNowSK TelecomSpaceXAIStepFunSwiss AITencentThinking MachinesTIITrillion LabsUpstageXiaomiZhipu AIOpen weightsNew modelsBest per model (431)431 of 431

#	Model	Index ↓	Provider ↕
1	GPT-6 AstramaxOpenAI	70.2	OpenAI
2	Claude Opus 5.5highAnthropic	70.1	Anthropic
3	Claude Fable 5.1xhighAnthropic	69.9	Anthropic
4	NewGemini 4 ArgonhighGoogle	69.7	Google
5	NewGPT-6.1 SolmaxOpenAI	69.5	OpenAI
6	Claude Fable 5thinkingAnthropic	69.2	Anthropic
7	Claude Opus 5highAnthropic	68.4	Anthropic
8	Muse Spark 1.3maxMeta	67.9	Meta
9	GPT-5.6 SolmaxOpenAI	67.5	OpenAI
10	NewClaude Sonnet 5.5xhighAnthropic	67.0	Anthropic
11	GPT-6 SolmaxOpenAI	66.3	OpenAI
12	GPT-5.5xhighOpenAI	65.9	OpenAI
13	Kimi K3maxMoonshot AIopen ↗	65.6	Moonshot AI
14	GPT-5.5 ProxhighOpenAI	65.6	OpenAI
15	Qwen3.8 MaxmaxAlibaba	65.0	Alibaba
16	MiMo-V2.6-ProXiaomiopen ↗	64.8	Xiaomi
17	Gemini 3.8 FlashmediumGoogle	64.7	Google
18	Muse SparkMeta	64.4	Meta
19	Grok 4.6mediumSpaceXAI	64.3	SpaceXAI
20	GPT-5.4xhighOpenAI	64.2	OpenAI
21	GLM 5.3maxZhipu AIopen ↗	64.1	Zhipu AI
22	GPT-5.3-CodexxhighOpenAI	64.0	OpenAI
23	Claude Opus 4.7Anthropic	63.8	Anthropic
24	GPT-5.6 TerramaxOpenAI	63.7	OpenAI
25	Grok 4.7highSpaceXAI	63.6	SpaceXAI
26	Muse Spark 1.1Meta	63.5	Meta
27	Qwen3.8 2.4T A95BAlibabaopen ↗	63.5	Alibaba
28	Gemini 3.7 FlashmediumGoogle	63.4	Google
29	GPT-5.4 ProOpenAI	63.4	OpenAI
30	Claude Opus 4.6Anthropic	63.1	Anthropic
31	Gemini 3.5 FlashmediumGoogle	63.1	Google
32	DeepSeek V4 PromaxDeepSeekopen ↗	62.9	DeepSeek
33	GLM 5.2maxZhipu AIopen ↗	62.7	Zhipu AI
34	Muse Spark 1.2xhighMeta	62.7	Meta
35	Gemini 3.1 ProGoogle	62.6	Google
36	Qwen3.6 MaxmaxAlibaba	62.5	Alibaba
37	Claude Opus 4.8maxAnthropic	62.4	Anthropic
38	GLM 5.3 FlashZhipu AIopen ↗	62.4	Zhipu AI
39	DeepSeek V4.1 FlashmaxDeepSeekopen ↗	61.7	DeepSeek
40	Step 5 PreviewStepFun	61.6	StepFun
41	GPT-5.2xhighOpenAI	60.9	OpenAI
42	Gemini 3 ProGoogle	60.8	Google
43	DeepSeek V4 FlashmaxDeepSeekopen ↗	60.7	DeepSeek
44	Claude Sonnet 5highAnthropic	60.6	Anthropic
45	Gemini 3.6 FlashhighGoogle	60.5	Google
46	Claude Opus 4.5thinkingAnthropic	60.2	Anthropic
47	GPT-5 ProOpenAI	60.2	OpenAI
48	Deepseek v4 Pro HighhighDeepSeek	60.1	DeepSeek
49	Qwen3 7maxAlibaba	59.9	Alibaba
50	Qwen3.8-Flash-NextAlibabaopen ↗	59.9	Alibaba
51	GPT-5.6 LunaxhighOpenAI	59.8	OpenAI
52	o3OpenAI	59.7	OpenAI
53	Grok 4.5highSpaceXAI	59.6	SpaceXAI
54	GPT-5.2-CodexxhighOpenAI	59.6	OpenAI
55	GPT-6 LunamaxOpenAI	59.4	OpenAI
56	MiMo-V2.6-FlashXiaomiopen ↗	59.3	Xiaomi
57	Gemini 3 FlashthinkingGoogle	59.3	Google
58	Grok 4.20thinkingSpaceXAI	59.1	SpaceXAI
59	GLM-5thinkingZhipu AIopen ↗	59.0	Zhipu AI
60	Mimo v2 ProXiaomi	58.9	Xiaomi

Show 100 more

Cite as: BenchLeader, “Quality: the BenchLeader Index”, https://www.benchleader.com/leaderboards/quality, data as of 5 Oct 2026.