Best LLM

What is the best LLM for coding?

Claude Fable 5.2, Claude Opus 5.5 and GPT-6 Astra are practically level at the top in coding.

Among the ten best, Claude Sonnet 5.5 is the cheapest at $4 per million tokens.

The fastest among them is Claude Sonnet 5.5, at 103 tokens per second.

Ranked models: 114
Tests: 17
Data as of: Nov 10, 2026

Ranking: Coding

Place and value in the area, plus overall index, price and speed. Higher values are better.

Rank	Model	Value	Overall index	Price (3:1)	Speed
1	Claude Fable 5.2Top group	131.6	130.9	$16	n/a
2	Claude Opus 5.5Top group	130.1	125.6	$8	75
3	Claude Sonnet 5.5Upper group	125.2	121.8	$4	103
4	GPT-6 AstraUpper group	124.3	120.2	$20	58
5	Gemini 4 CarbonUpper group	123.5	127.0	n/a	n/a
6	Claude Fable 5.1Upper group	122.8	124.1	$20	59
7	Gemini 4 ArgonUpper group	122.1	121.4	$4	46
8	GPT-6.1 SolUpper group	120.5	118.0	$4	44
9	Claude Opus 5Upper group	118.4	117.4	$10	71
10	Kimi K3.1 ProUpper group	116.6	112.6	$4	56
11	Muse Wonder FullUpper group	116.1	118.1	n/a	n/a
12	GPT-6 SolUpper group	115.2	109.1	$4	82
13	GPT-6.1 TerraMiddle group	114.7	106.5	$2	62
14	Claude Fable 5Middle group	113.2	112.8	$20	50
15	Muse WonderMiddle group	112.9	113.3	$8	n/a
16	GPT-5.6 SolMiddle group	111.9	109.1	$8	75
17	Muse Ember 1.0Middle group	111.0	110.4	$0.20	n/a
18	DeepSeek V4.1 Flash 1017Middle group	109.4	106.7	$0.26	126
19	Hy4Middle group	109.2	n/a	n/a	n/a
20	Claude Haiku 5.5Middle group	106.3	102.8	$0.20	164

Show all 114 ranked models

Rank	Model	Value	Overall index	Price (3:1)	Speed
21	Muse Spark 1.3Middle group	106.0	107.3	$2	102
22	GLM-5.3Middle group	105.4	100.6	$2.15	81
23	GPT-5.6 TerraMiddle group	104.5	100.8	$4.50	72
24	Grok 4.7Middle group	103.3	103.3	$3	47
25	Mistral Large 4Middle group	102.0	100.8	$1.05	n/a
26	Claude Opus 4.8Middle group	101.5	99.5	$10	71
27	MiniMax M3.1 FlashMiddle group	101.3	100.0	$0.26	n/a
28	Qwen 3.8 Max 0902Middle group	101.0	100.8	$2.48	36
29	Grok 4.6Middle group	100.9	102.1	$3	58
30	DeepSeek V4 Pro 0813Middle group	100.5	n/a	n/a	n/a
31	Gemini 3.8 FlashMiddle group	99.9	106.8	$1.50	81
32	MiMo V2.6 ProMiddle group	99.8	100.4	$0.54	32
33	Ember-1Middle group	99.2	n/a	n/a	n/a
34	DeepSeek-V4.1-FlashMiddle group	98.8	99.5	$0.26	118
35	Ling 3.1 Flash 1019Middle group	98.3	n/a	n/a	n/a
36	MiMo V2.6 FlashMiddle group	98.2	97.3	$0.18	38
37	Step 5 PreviewMiddle group	97.7	97.1	$1.43	65
38	Inkling 2 SmallMiddle group	97.7	n/a	n/a	n/a
39	GPT-6 LunaMiddle group	97.6	95.8	$0.20	77
40	Gemini 3.7 FlashMiddle group	97.6	107.9	n/a	72
41	Hy4 previewMiddle group	97.3	n/a	n/a	n/a
42	Kimi K3Middle group	97.3	101.1	$6	61
43	GLM-5.3-FlashMiddle group	96.9	94.3	$0.24	101
44	GPT-5.6 LunaMiddle group	96.5	93.8	$0.45	123
45	DeepSeek V4 Flash 0731Middle group	95.4	n/a	n/a	n/a
46	GPT-5.5Middle group	95.1	101.5	$11.25	87
47	Claude Sonnet 5Middle group	94.1	92.1	$4	77
48	Claude Opus 4.7Middle group	93.7	95.1	$10	57
49	Grok 4.5Middle group	92.9	94.8	$3	52
50	GPT-5.3-CodexMiddle group	91.7	n/a	$4.81	n/a
51	Gemini 3.6 FlashMiddle group	91.3	97.6	$1.50	110
52	DeepSeek-V4-ProMiddle group	91.0	87.2	$0.99	44
53	Agnes 2.5 Pro ThetaMiddle group	90.8	n/a	$0.15	n/a
54	Gemini 3.5 FlashMiddle group	89.8	97.8	n/a	142
55	GLM-5.2Middle group	89.5	89.0	$2.15	73
56	Muse Spark OpenMiddle group	88.3	97.5	n/a	n/a
57	Mercury 3Middle group	87.7	n/a	$0.16	n/a
58	Claude Sonnet 4.6Middle group	87.2	86.7	$6	38
59	GPT-5.4Middle group	85.1	86.5	$5.63	58
60	Kimi K2.6Lower group	83.9	84.9	$1.71	32
61	Qwen 3.8 27BLower group	83.4	82.8	$0.66	66
62	Kimi K2.7 CodeLower group	82.2	81.3	$1.71	50
63	Reflection BeamLower group	81.8	82.4	n/a	n/a
64	GLM-5.1Lower group	81.2	82.8	$2.15	39
65	Gemini 3.1 Pro PreviewLower group	80.9	91.4	$4.50	96
66	Qwen 3.7 PlusLower group	80.7	83.4	$0.48	50
67	MiniMax M3Lower group	79.9	83.0	$0.52	82
68	Qwen 3.7 MaxLower group	79.4	85.4	$2.48	50
69	K2 Horizon 375B-A23BLower group	78.5	n/a	n/a	n/a
70	GPT-5.4 miniLower group	75.7	75.1	$1.69	134
71	o3Lower group	75.6	74.2	$3.50	73
72	MiMo-V2.5Lower group	74.6	69.1	$0.18	55
73	GPT-5Lower group	74.0	77.5	$3.44	83
74	MiMo-V2.5-ProLower group	73.4	73.0	$0.54	59
75	GPT-5.4 nanoLower group	72.4	64.0	$0.46	98
76	Gemini 3 Flash PreviewLower group	71.8	81.0	$1.13	101
77	GLM-5Lower group	71.5	72.7	$1.55	41
78	o4-miniLower group	69.9	63.5	n/a	n/a
79	InklingLower group	69.5	72.1	$1.76	n/a
80	Gemini 3.5 Flash-LiteLower group	68.9	73.7	$0.85	145
81	Kimi K2.5Lower group	67.7	70.3	$1.14	37
82	Grok 4.3Lower group	67.4	71.6	$1.56	66
83	MiniMax M2.7Lower group	67.0	65.1	$0.37	45
84	MiniMax M2.1Lower group	66.6	57.3	$0.52	n/a
85	GLM-4.7Lower group	66.5	56.2	$0.78	25
86	GPT-5 miniLower group	65.6	65.8	$0.69	105
87	MiniMax M2.5Lower group	65.5	61.5	$0.52	49
88	Nemotron 3.5 Super VL 120B A12BLower group	65.5	68.1	$0.24	n/a
89	Inkling SmallLower group	64.5	68.7	$0.52	61
90	Nemotron 3 UltraLower group	60.1	66.8	$0.93	92
91	Mercury 2.5Lower group	58.7	45.3	n/a	n/a
92	Muse Glimmer 30BLower group	58.3	n/a	$0.52	124
93	GLM-4.6Lower group	51.2	45.9	$0.88	28
94	DeepSeek-V3.2Lower group	50.8	48.5	$0.29	31
95	Mistral Medium 3.5Lower group	50.1	46.3	$3	118
96	Kimi K2 ThinkingLower group	48.7	53.2	$1.08	43
97	Gemini 3.1 Flash-LiteLower group	47.7	55.9	$0.56	125
98	Grok 4.20 ReasoningLower group	46.8	58.3	$1.56	66
99	Claude Haiku 4.5Lower group	46.5	54.1	$2	84
100	Laguna XS.2Lower group	45.6	n/a	n/a	n/a
101	GPT-4.1Lower group	45.5	43.6	$3.50	81
102	Ling 3.0 FlashLower group	42.6	47.9	$0.032	106
103	DeepSeek-R1Lower group	42.0	44.5	$1.15	n/a
104	Nemotron 3.5 LightningLower group	41.0	31.1	$0.085	180
105	North Mini CodeLower group	40.3	n/a	n/a	n/a
106	Granite 4.2 8BLower group	40.3	n/a	n/a	n/a
107	Gemini 2.5 ProLower group	39.6	61.4	$3.44	74
108	Granite 4.2 3BLower group	38.1	n/a	n/a	n/a
109	GPT OSS 120BLower group	38.1	43.4	$0.20	92
110	Command A PlusLower group	33.9	30.7	n/a	n/a
111	Mistral Large 3Lower group	32.0	32.7	$0.75	86
112	Mistral Small 4Lower group	31.4	23.3	$0.26	112
113	DeepSeek-V3Lower group	16.7	19.7	$0.37	40
114	Command ALower group	-14.8	0.4	$4.38	2

Price in dollars per million tokens, mixed from 3 parts input and 1 part output. Speed in tokens per second.

The top 20: CodingValue in the area · higher is better

Claude Fable 5.2Rank 1131.6

Claude Opus 5.5Rank 2130.1

Claude Sonnet 5.5Rank 3125.2

GPT-6 AstraRank 4124.3

Gemini 4 CarbonRank 5123.5

Claude Fable 5.1Rank 6122.8

Gemini 4 ArgonRank 7122.1

GPT-6.1 SolRank 8120.5

Claude Opus 5Rank 9118.4

Kimi K3.1 ProRank 10116.6

Muse Wonder FullRank 11116.1

GPT-6 SolRank 12115.2

GPT-6.1 TerraRank 13114.7

Claude Fable 5Rank 14113.2

Muse WonderRank 15112.9

GPT-5.6 SolRank 16111.9

Muse Ember 1.0Rank 17111.0

DeepSeek V4.1 Flash 1017Rank 18109.4

Hy4Rank 19109.2

Claude Haiku 5.5Rank 20106.3

Which tests count here?

17 tests by independent evaluators feed this area. Each test counts with the result of the model in question.

Vibe Code Bench v1.1
95 models tested · Vals AI
LiveCodeBench
84 models tested · Vals AI
Terminal-Bench
76 models tested · Artificial Analysis, Vals AI
Code Migration
71 models tested · Vals AI
Terminal-Bench Science
71 models tested · Artificial Analysis, Vals AI
SWE-bench Verified
68 models tested · Vals AI, Princeton HAL
LMArena Code Arena WebDev
61 models tested · LMArena
ProgramBench
59 models tested · Vals AI
FrontierCode
57 models tested · Artificial Analysis
IOI
43 models tested · Vals AI
FrontierCode 1.1 Main
40 models tested · Artificial Analysis
Code Migration v1.1
29 models tested · Vals AI
Vibe Code Bench 1-100
24 models tested · Vals AI
SWE-bench Multimodal
21 models tested · Vals AI
ScienceAgentBench
6 models tested · Princeton HAL
USACO
6 models tested · Princeton HAL
SciCode
5 models tested · Princeton HAL

Popular comparisons among the best

Other tasks

All LLMs compared How the overall index works Compare several models