Best AI Models for Coding in 2026

Whether you're building apps, debugging, or writing scripts, these models score highest on coding benchmarks. Ranked by real coding evaluation data from Artificial Analysis.

#		Model	Provider	Intelligence	Coding	Speed	Price/1M
1		Google: Gemini 3.1 Pro Preview	Google	57.2	55.5	117 tok/s	$2.00
2		Google: Gemini 3.1 Pro Preview Custom Tools	Google	57.2	55.5	117 tok/s	$2.00
3		Anthropic: Claude Opus 4	Anthropic	46.5	47.6	44 tok/s	$15
4		Google: Nano Banana Pro (Gemini 3 Pro Image Preview)	Google	48.4	46.5	119 tok/s	$2.00
5		Anthropic: Claude Sonnet 4	Anthropic	44.4	46.4	44 tok/s	$3.00
6		OpenAI: GPT-5	OpenAI	44.4	43.9	221 tok/s	$1.25
7		MiniMax: MiniMax M2	MiniMax	49.6	41.9	44 tok/s	$0.26
8		Google: Gemini 3 Pro Preview	Google	41.3	39.4	116 tok/s	$2.00
9		OpenAI: GPT-5 Codex	OpenAI	44.6	38.9	170 tok/s	$1.25
10		OpenAI: o3 Deep Research	OpenAI	38.4	38.4	77 tok/s	$10
11		OpenAI: o3 Pro	OpenAI	38.4	38.4	77 tok/s	$20
12		Google: Gemini 3 Flash Preview	Google	35.0	37.8	186 tok/s	$0.50
13		DeepSeek: DeepSeek V3.2	DeepSeek	41.7	36.7	36 tok/s	$0.26
14		DeepSeek: DeepSeek V3.2 Speciale	DeepSeek	41.7	36.7	36 tok/s	$0.40
15		OpenAI: GPT-5.4 Nano	OpenAI	44.6	36.0	61 tok/s	$0.20
16		OpenAI: GPT-5.4 Mini	OpenAI	44.6	36.0	61 tok/s	$0.75
17		OpenAI: o1	OpenAI	23.7	34.0	0 tok/s	$15
18		Anthropic: Claude 3 Haiku	Anthropic	37.1	32.6	118 tok/s	$0.25
19		Anthropic: Claude Haiku 4.5	Anthropic	37.1	32.6	118 tok/s	$1.00
20		Google: Gemini 2.5 Pro	Google	34.6	31.9	120 tok/s	$1.25
21		Google: Gemini 2.5 Pro Preview 06-05	Google	34.6	31.9	120 tok/s	$1.25
22	S	StepFun: Step 3.5 Flash	StepFun	37.8	31.6	78 tok/s	$0.10
23		NVIDIA: Nemotron 3 Super	NVIDIA	36.0	31.2	402 tok/s	$0.10
24	A	Amazon: Nova Pro 1.0	Amazon	35.7	30.4	151 tok/s	$0.80
25		Anthropic: Claude 3.5 Sonnet	Anthropic	15.9	30.2	0 tok/s	$6.00
26		Google: Gemini 2.5 Flash Lite Preview 09-2025	Google	33.5	30.1	214 tok/s	$0.10
27		Google: Gemini 3.1 Flash Lite Preview	Google	33.5	30.1	214 tok/s	$0.25
28		MiniMax: MiniMax M2.5	MiniMax	36.1	29.2	46 tok/s	$0.20
29		MiniMax: MiniMax M2.7	MiniMax	36.1	29.2	46 tok/s	$0.30
30		OpenAI: gpt-oss-120b	OpenAI	33.3	28.6	289 tok/s	$0.04
31		DeepSeek: DeepSeek V3.1 Terminus	DeepSeek	28.1	28.4	0 tok/s	$0.21
32		Anthropic: Claude 3.7 Sonnet	Anthropic	34.7	27.6	0 tok/s	$3.00
33		Anthropic: Claude 3.7 Sonnet (thinking)	Anthropic	34.7	27.6	0 tok/s	$3.00
34		xAI: Grok 4 Fast	xAI	35.1	27.4	137 tok/s	$0.20
35		OpenAI: o4 Mini Deep Research	OpenAI	33.1	25.6	140 tok/s	$2.00
36		xAI: Grok 4	xAI	29.7	25.4	87 tok/s	$3.00
37		xAI: Grok 3 Mini	xAI	32.1	25.2	199 tok/s	$0.30
38		Mistral: Mistral Small 4	Mistral AI	26.9	24.3	0 tok/s	$0.15
39		DeepSeek: R1 0528	DeepSeek	27.1	24.0	0 tok/s	$0.45
40		xAI: Grok Code Fast 1	xAI	28.7	23.7	172 tok/s	$0.20
41		Mistral: Devstral 2 2512	Mistral AI	22.0	23.7	79 tok/s	$0.40
42		Mistral Large	Mistral AI	22.8	22.7	60 tok/s	$2.00
43		DeepSeek: DeepSeek V3 0324	DeepSeek	22.3	22.0	0 tok/s	$0.20
44		OpenAI: GPT-5 Mini	OpenAI	20.7	21.9	92 tok/s	$0.25
45		OpenAI: GPT-5 Chat	OpenAI	21.8	21.2	116 tok/s	$1.25
46		Mistral: Devstral Small 1.1	Mistral AI	19.5	20.7	204 tok/s	$0.10
47		OpenAI: o1-pro	OpenAI	30.8	20.5	115 tok/s	$150
48		Anthropic: Claude Opus 4.5	Anthropic	18.0	19.5	0 tok/s	$5.00
49		Anthropic: Claude Opus 4.1	Anthropic	18.0	19.5	0 tok/s	$15
50		Anthropic: Claude Opus 4.6	Anthropic	18.0	19.5	0 tok/s	$5.00

Benchmark data by Artificial Analysis