Scores from Oct 28, 2026 · refreshed daily

The best AI models, ranked.

67 models graded on public benchmarks for everyday questions, web apps and code, then weighed against what they cost you. One key calls every one that’s available.

Best overall

Claude Opus 5.5by Anthropic · Coming soon

The highest overall grade of any model we grade.

Price at launch$8.00 / 1M

Best value

DeepSeek V4 Flashby DeepSeek · Coming soon

The most grade for every dollar, among models rated Strong or better.

Price at launch$0.11 / 1M

Best budget pick

MiMo V2.6 Flashby Xiaomi · Coming soon

The lowest price of any model rated Excellent.

Price at launch$0.17 / 1M

Overall grade against price

Up and to the left is better. The line joins the models nothing else beats on both grade and price.

Blended price per 1M tokens (3 input : 1 output), log scale

Leaderboard

74 models · ranked by overall grade

Sort

#	Model	Context	Price in / outPrice
1	Claude Opus 5.5Anthropic · Coming soon	1M	$4.00 / $20.00$8.00
2	Claude Sonnet 5.5Anthropic · Coming soon	1M	$2.00 / $10.00$4.00
3	Claude Fable 5Anthropic · Coming soon	1M	$10.00 / $50.00$20.00
4	Claude Opus 5Anthropic · Coming soon	1M	$5.00 / $25.00$10.00
5	Kimi K3Moonshot AI	1M	$2.85 / $14.25$5.70
6	Claude Opus 4.7Anthropic · Coming soon	1M	$5.00 / $25.00$10.00
7	Qwen3.8 MaxQwen · Coming soon	250K	$1.65 / $4.95$2.48
8	MiMo V2.6 ProXiaomi · Coming soon	1M	$0.43 / $0.87$0.54
9	Seed 2.0 ProByteDance · Coming soon	250K	$0.50 / $3.00$1.13
10	GLM 5.3Z.ai · Coming soon	1M	$0.90 / $4.00$1.68
11	DeepSeek V4.1 FlashDeepSeek · Coming soon	1M	$0.20 / $0.60$0.30
11	Gemini 3.7 FlashGoogle · Coming soon	1M	$0.75 / $3.75$1.50
13	GLM 5.3 FlashZ.ai · Coming soon	1M	$0.15 / $0.50$0.24
14	Claude Opus 4.8Anthropic · Coming soon	1M	$5.00 / $25.00$10.00
15	GLM 5.2Z.ai · Coming soon	1M	$0.75 / $2.40$1.16
16	MiMo V2.6 FlashXiaomi · Coming soon	1M	$0.14 / $0.28$0.17
17	Claude Sonnet 4.6Anthropic · Coming soon	1M	$3.00 / $15.00$6.00
18	Qwen3.7 MaxQwen · Coming soon	250K	$2.50 / $7.50$3.75
19	DeepSeek V4 Pro 0813DeepSeek · Coming soon	1M	$1.30 / $2.60$1.63
20	Claude Sonnet 5Anthropic · Coming soon	1M	$3.00 / $15.00$6.00
20	Gemini 3.1 ProGoogle · Coming soon	1M	$2.00 / $12.00$4.50
22	Gemini 3.5 FlashGoogle · Coming soon	1M	$1.50 / $9.00$3.38
23	Kimi K2.6Moonshot AI · Coming soon	256K	$0.75 / $3.50$1.44
24	Step 5 PreviewStepFun · Coming soon	1M	$1.00 / $2.70$1.43
25	Qwen3.8 27BQwen · Coming soon	256K	$0.20 / $2.50$0.78
26	NVIDIA Nemotron 3 Ultra 550B A55BNVIDIA · Coming soon	256K	$0.50 / $2.20$0.93
27	DeepSeek V4 FlashDeepSeek · Coming soon	1M	$0.090 / $0.18$0.11
28	Qwen3 235B A22B Instruct 2507Qwen · Coming soon	256K	$0.090 / $0.55$0.20
29	DeepSeek R1 0528DeepSeek · Coming soon	160K	$0.50 / $2.15$0.91
30	MiniMax M3MiniMax · Coming soon	512K	$0.28 / $1.10$0.48
31	DeepSeek V3.1DeepSeek · Coming soon	160K	$0.25 / $0.95$0.42
31	Qwen3 VL 235B A22B InstructQwen · Coming soon	256K	$0.20 / $0.88$0.37
33	Gemma 4 31B ITGoogle · Coming soon	256K	$0.20 / $0.40$0.25
34	GLM 4.7Z.ai · Coming soon	198K	$0.40 / $1.75$0.74
35	Qwen3.5 397B A17BQwen · Coming soon	256K	$0.45 / $3.00$1.09
36	Gemini 2.5 FlashGoogle · Coming soon	1M	$0.30 / $2.50$0.85
37	Qwen3 Next 80B A3B InstructQwen · Coming soon	256K	$0.090 / $1.10$0.34
38	Gemma 4 26B A4B ITGoogle · Coming soon	256K	$0.070 / $0.34$0.14
39	Muse Glimmer 30BMeta · Coming soon	128K	$0.30 / $1.20$0.53
40	DeepSeek V3.2DeepSeek · Coming soon	160K	$0.26 / $0.38$0.29
41	Inkling SmallComing soon	512K	$0.45 / $1.20$0.64
42	Claude Haiku 4.5Anthropic · Coming soon	195K	$1.00 / $5.00$2.00
43	GLM 4.6Z.ai · Coming soon	198K	$0.50 / $2.00$0.88
44	Gemini 2.5 ProGoogle · Coming soon	1M	$1.25 / $10.00$3.44
45	Qwen3.5 27BQwen · Coming soon	256K	$0.26 / $2.60$0.84
46	Gemini 3.1 Flash LiteGoogle · Coming soon	1M	$0.25 / $1.50$0.56
47	NVIDIA Nemotron 3.5 LightningNVIDIA · Coming soon	256K	$0.060 / $0.16$0.085
48	NVIDIA Nemotron 3 Super 120B A12BNVIDIA · Coming soon	256K	$0.075 / $0.40$0.16
49	Mistral Small 3.2 24B Instruct 2506Mistral AI · Coming soon	125K	$0.075 / $0.20$0.11
50	Qwen3 Coder 480B A35B Instruct TurboQwen · Coming soon	256K	$0.30 / $1.00$0.47
51	DeepSeek V3DeepSeek · Coming soon	160K	$0.32 / $0.89$0.46
52	Qwen3 32BQwen · Coming soon	40K	$0.080 / $0.28$0.13
53	GPT OSS 120BOpenAI · Coming soon	128K	$0.037 / $0.17$0.070
54	Qwen3.5 35B A3BQwen · Coming soon	256K	$0.14 / $1.00$0.35
55	Gemma 3 27B ITGoogle · Coming soon	128K	$0.080 / $0.16$0.10
56	Qwen3 30B A3BQwen · Coming soon	40K	$0.12 / $0.50$0.21
57	Llama 4 Scout 17B 16E InstructMeta · Coming soon	320K	$0.10 / $0.30$0.15
58	GPT OSS 20BOpenAI · Coming soon	128K	$0.030 / $0.14$0.058
59	Gemma 3 12B ITGoogle · Coming soon	128K	$0.050 / $0.15$0.075
60	Llama 3.3 70B Instruct TurboMeta · Coming soon	128K	$0.10 / $0.32$0.15
61	Qwen2.5 72B InstructQwen · Coming soon	32K	$0.36 / $0.40$0.37
62	Granite 4.2 8BIBM · Coming soon	128K	$0.060 / $0.25$0.11
63	Meta Llama 3.1 70B Instruct TurboMeta · Coming soon	128K	$0.40 / $0.40$0.40
64	Gemma 3 4B ITGoogle · Coming soon	128K	$0.050 / $0.10$0.063
65	Mistral Small 24B Instruct 2501Mistral AI · Coming soon	32K	$0.050 / $0.080$0.058
66	Phi 4Microsoft · Coming soon	16K	$0.070 / $0.14$0.087
67	Meta Llama 3.1 8B Instruct TurboMeta · Coming soon	128K	$0.020 / $0.040$0.025
–	Hy4 PreviewComing soon	1M	$0.83 / $2.50$1.25
–	DeepSeek V4 Flash 0731DeepSeek · Coming soon	1M	$0.060 / $0.18$0.090
–	Ling 3.1 FlashinclusionAI · Coming soon	256K	$0.30 / $0.90$0.45
–	Qwen3.5 9BQwen · Coming soon	256K	$0.10 / $0.15$0.11
–	Qwen3.6 27BQwen · Coming soon	256K	$0.32 / $3.20$1.04
–	Qwen3.6 35B A3BQwen · Coming soon	256K	$0.10 / $0.95$0.31
–	Qwen3.8 FlashQwen · Coming soon	1M	$0.11 / $0.38$0.18

Prices per million tokens, input / output; for models coming soon, the price at launch.Blended price per million tokens (3 input : 1 output); at launch for models coming soon.

How the grade works

Each benchmark is put on a 0–100 scale against every model we grade, then weighted into one overall grade. A model needs scores covering 40% of the weight to be rated; otherwise it shows as Not rated. Arena ratings are relative and dated, so the grade moves as new models arrive.

Excellent · 70+Strong · 55–70Good · under 55Not rated

LMArena Textweight 35 · LMArena (CC BY 4.0)
LMArena WebDevweight 20 · LMArena (CC BY 4.0)
LMArena Codingweight 20 · LMArena (CC BY 4.0)
Aider Polyglotweight 3 · Aider (aider.chat)
SWE-bench Verified, SWE-Bench Pro, Terminal-Bench 4.0, Terminal-Bench 3.0, Terminal-Bench 2.1, Terminal-Bench 2.0, LiveCodeBench are shown for reference on the benchmarks page; they don’t count toward the grade.

Every model on this page. One key.

Works in the tools you already use, and every model is hosted in the USA. Switch models by changing one word, and pay only for the tokens you use. Models marked coming soon are priced as they’ll launch.

Get your API key Read the quickstart

Scores: LMArena (CC BY 4.0), Aider (aider.chat), SWE-bench, Scale AI SWE-Bench Pro, Terminal-Bench, LiveCodeBench, and model makers’ published results where marked v. Prices are FullsendAI’s, per million tokens. Data as of Oct 28, 2026. The same grades as JSON: the public benchmarks API.