Best for agents: the full table

Highest agents-and-tools category score, among models with at least three agentic results.

Looking for the short answer? Best LLM for agents (October 2026) has the top pick, best value and open-weights options. This page is every qualifying model, ranked.

As of 24 Oct 2026, Claude Sonnet 5.5 by Anthropic leads this list (69.7), ahead of Kimi K3.1 Pro (67.3). Best LLM for agents (October 2026)

64 models qualify · data as of 24 Oct 2026

1. 1Claude Sonnet 5.5 (max)69.7
2. 2Kimi K3.1 Pro67.3
3. 3GPT-5.6 Sol (max)66.5
4. 4Muse Spark 1.3 (max)65.6
5. 5Claude Opus 4.765.4
6. 6Claude Fable 5 (max)65
7. 7GPT-6.1 Sol (max)64.6
8. 8GPT-6 Astra (max)64.2
9. 9Claude Opus 5 (max)64.1
10. 10GPT-5.4 (xhigh)63.6
11. 11Gemini 3.5 Flash (high)63.4
12. 12Claude Opus 5.5 (max)63
13. 13DeepSeek V4.1 Flash 1017 (max)63
14. 14GLM 5.2 (max)62.6
15. 15Gemini 3 Pro61.8

SearchAll providersAlibabaAnthropicDeepSeekGoogleMetaMiniMaxMistral AIMoonshot AINVIDIAOpenAIReflectionSpaceXAIStepFunThinking MachinesXiaomiZhipu AIOpen weightsNew models64 of 64

#	Model	Agents ↓	Index ↕	Price $/M ↕	Speed tok/s ↕
1	Claude Sonnet 5.5maxAnthropic	69.7	69.3	$4.00	133
2	NewKimi K3.1 ProMoonshot AI	67.3	66.8	$4.00	81
3	GPT-5.6 SolmaxOpenAI	66.5	68.0	$8.00	86
4	Muse Spark 1.3maxMeta	65.6	67.1	$2.00	116
5	Claude Opus 4.7Anthropic	65.4	63.3	$10.00	60
6	Claude Fable 5maxAnthropic	65	69.3	$20.00	70
7	GPT-6.1 SolmaxOpenAI	64.6	69.9	$4.00	55
8	GPT-6 AstramaxOpenAI	64.2	70.4	$20.00	58
9	Claude Opus 5maxAnthropic	64.1	68.8	$10.00	57
10	GPT-5.4xhighOpenAI	63.6	64.7	$5.63	96
11	Gemini 3.5 FlashhighGoogle	63.4	63.1	$3.38	222
12	Claude Opus 5.5maxAnthropic	63	72.6	$8.00	96
13	NewDeepSeek V4.1 Flash 1017maxDeepSeekopen ↗	63	64.0	$0.525	198
14	GLM 5.2maxZhipu AIopen ↗	62.6	62.3	$2.15	91
15	Gemini 3 ProGoogle	61.8	60.4	–	–
16	GPT-5.5xhighOpenAI	61.2	66.1	$11.25	86
17	Claude Fable 5.1maxAnthropic	60.2	70.5	$20.00	68
18	GPT-6 SolmaxOpenAI	59.7	66.4	$4.00	103
19	GLM 5.3 FlashZhipu AIopen ↗	59.4	62.6	$0.238	69
20	Kimi K3maxMoonshot AIopen ↗	58.9	64.7	$6.00	39
21	GPT-5.6 TerramaxOpenAI	58.9	64.4	$4.50	126
22	MiMo-V2.6-ProXiaomiopen ↗	58.3	64.0	$0.548	54
23	Qwen3.8 MaxmaxAlibaba	58.3	63.3	$3.00	43
24	Grok 4SpaceXAI	57.6	58.0	$6.00	78
25	GLM 5.3maxZhipu AIopen ↗	57.5	63.5	$2.15	85
26	Gemini 3.1 ProGoogle	57.1	60.6	$4.50	132
27	DeepSeek V4.1 FlashmaxDeepSeekopen ↗	56.6	61.2	$0.525	181
28	Claude Opus 4.8maxAnthropic	56.5	63.3	$10.00	63
29	Qwen3.5 397B A17BthinkingAlibabaopen ↗	56.2	55.1	$1.35	57
30	Muse Spark 1.2xhighMeta	55	63.2	$2.00	307
31	Qwen3.6 PlusAlibaba	53.1	58.8	$1.13	56
32	Kimi K2Moonshot AIopen ↗	52.4	50.4	$1.00	47
33	Kimi K2.7 CodeMoonshot AIopen ↗	52.2	55.8	$1.71	54
34	Muse Spark 1.1Meta	51.5	62.5	$2.00	200
35	Claude Sonnet 4.6Anthropic	51.3	57.1	$6.00	44
36	Gemini 3.8 FlashhighGoogle	50.9	64.4	$1.50	141
37	GPT-5.4 MinixhighOpenAI	50.2	56.3	$1.69	241
38	Step 5 PreviewStepFunopen	50.1	60.4	$1.43	224
39	Grok 4.5highSpaceXAI	50.1	59.3	$3.00	72
40	GPT-5.6 LunamaxOpenAI	49	59.9	$0.450	129
41	Mistral Large 4Mistral AIopen	48.9	56.6	$2.06	93
42	Qwen3.7maxAlibaba	48.7	60.6	$3.75	201
43	GPT-6 LunamaxOpenAI	48.6	60.0	$0.200	136
44	Gemini 3.7 FlashhighGoogle	47.5	63.5	$1.50	88
45	MiniMax M2.7MiniMaxopen ↗	47.4	55.6	$0.525	55
46	Qwen3.8 27BxhighAlibabaopen ↗	47.3	56.2	$1.13	45
47	MiniMax M2.5MiniMaxopen ↗	46	53.6	$0.525	68
48	TInkling SmallThinking Machinesopen ↗	45.6	54.8	$0.525	224
49	Qwen3.7 PlusAlibaba	45.2	58.5	$0.700	59
50	Kimi K2.5thinkingMoonshot AIopen ↗	45.1	57.6	$1.20	66
51	Mistral Medium 3.5Mistral AIopen	43.8	49.4	$3.00	159
52	GPT-5.1highOpenAI	41.8	57.7	$3.44	111
53	NewBeam 1.0Reflectionopen	41.7	51.1	$0.600	337
54	Gemini 2.5 ProGoogle	41.2	53.9	$3.44	121
55	Mistral Small 3.2Mistral AIopen ↗	41	41.6	$0.110	42
56	MiniMax M3MiniMaxopen ↗	40.1	56.4	$0.525	84
57	Llama 4 MaverickMetaopen ↗	40	42.8	$0.418	49
58	GPT-4.1 nanoOpenAI	37.9	38.0	$0.175	223
59	Nemotron 3 Ultra 550B A55BthinkingNVIDIAopen ↗	37.3	55.5	$1.13	159
60	Llama 4 ScoutMetaopen ↗	35.6	39.7	$0.300	105

Show 4 more