Neura Intelligence Index

AI Model Benchmarks

Q: What is the Neura Intelligence Index?

The Neura Intelligence Index is a composite score (0-100) that aggregates AI model performance across 15+ benchmarks spanning coding, math, reasoning, general knowledge, language, multimodal, safety, and agentic tasks. It uses min-max normalization and weighted category averaging to produce a single comparable score.

Q: How often are benchmark scores updated?

Benchmark scores are synced daily via an automated pipeline that aggregates data from official model cards, Papers With Code, HuggingFace Open LLM Leaderboard, LiveBench, and LMSYS Chatbot Arena.

Q: What does the confidence level mean?

Confidence reflects how many benchmarks a model has been tested on relative to the total. High means >70% benchmark coverage, medium is 40-70%, and low is <40%. Models with low coverage may have composite scores that shift as more benchmarks are added.

Q: How are category weights determined?

Category weights reflect real-world demand: Coding and Reasoning each get 20%, General and Math each get 15%, Language and Multimodal each get 10%, and Safety and Agent each get 5%. Weights are renormalized across available categories so models are not penalized for missing data.

Q: Can I compare specific models side by side?

Yes. Click any model to see its full benchmark profile, or use the comparison feature to compare up to 4 models side by side with radar charts and benchmark-by-benchmark scoring.

Compare AI model performance across 48+ benchmarks. Our composite index aggregates coding, math, reasoning, and language scores into a single intelligence ranking.

372

Models Ranked

Benchmarks

Jul 21, 2026

Last Updated

All(372)LLMs(350)Image(10)Video(8)Agent(4)

Highlights

At-a-glance rankings across the three dimensions that matter most.

Intelligence

Neura Intelligence Index; Higher is better

Context Window

Max input tokens; Higher is better

Price

USD per 1M output tokens; Lower is better

Leaderboard

Ranked by the Neura Intelligence Index — a weighted composite of 48 benchmarks across 8 categories.

Intelligence Tradeoffs

Find the sweet spot — models in the top-left quadrant offer the best value.

Intelligence vs. Output Cost ($/1M tokens)

Google

DeepSeek

Zhipu AI

StepFun

Alibaba Cloud / Qwen Team

OpenAI

Amazon

Moonshot AI

Anthropic

Category Leaders

Top performers in each benchmark category.

Top Models per Task

Who leads on each individual benchmark — click any card to see full results.

Best in GPQA

GPQA Diamond (%)

Best in SWE-bench

SWE-bench Verified (%)

Best in MATH

MATH (%)

Best in MMLU-Pro

MMLU-Pro (%)

Best in HumanEval

HumanEval (%)

Best in Chatbot Arena Elo

Chatbot Arena Elo (Elo)

Best Model For Your Use Case

Different tasks need different strengths. These indices re-weight our benchmarks for specific workflows.

Coding & Engineering

Best for software development, code generation, and debugging

Claude Fable 597.7 2

Claude Mythos Preview94.1 3

Claude Opus 4.892.9 4

Kimi K391.8 5

GPT-5.6 Sol89.9

Research & Analysis

Best for scientific research, data analysis, and complex reasoning

Content & Copywriting

Best for writing, editing, summarization, and creative tasks

Best Value

Highest intelligence per dollar — the cost-efficiency sweet spot

Step-3.5-Flash182.4 2

Methodology

The Neura Intelligence Index is a composite score (0-100) computed from 48+ individual benchmarks spanning 8 categories: coding, math, reasoning, general knowledge, language, multimodal, safety, and agentic tasks.

How it works

Normalize — Each benchmark score is min-max normalized to 0-1 across all models.
Categorize — Normalized scores are averaged within their benchmark category.
Weight — Category averages are combined using weights: Coding (20%), Reasoning (20%), General (15%), Math (15%), Language (10%), Multimodal (10%), Safety (5%), Agent (5%).
Scale — The weighted average is scaled to 0-100.

Coverage & confidence

Not all models have scores on all benchmarks. The confidence indicator reflects benchmark coverage: high (>70% of benchmarks), medium (40-70%), or low (<40%). Weights are renormalized across available categories so models aren't penalized for missing data.

Data sources

Scores are aggregated from official model cards, Papers With Code, HuggingFace Open LLM Leaderboard, LiveBench, and LMSYS Chatbot Arena. Each score includes a verification status (official, self-reported, or aggregated).

Frequently Asked Questions

What is the Neura Intelligence Index?▾

The Neura Intelligence Index is a composite score (0-100) that aggregates AI model performance across 15+ benchmarks spanning coding, math, reasoning, general knowledge, language, multimodal, safety, and agentic tasks. It uses min-max normalization and weighted category averaging to produce a single comparable score.

How often are benchmark scores updated?▾

Benchmark scores are synced daily via an automated pipeline that aggregates data from official model cards, Papers With Code, HuggingFace Open LLM Leaderboard, LiveBench, and LMSYS Chatbot Arena.

What does the confidence level mean?▾

Confidence reflects how many benchmarks a model has been tested on relative to the total. High means >70% benchmark coverage, medium is 40-70%, and low is <40%. Models with low coverage may have composite scores that shift as more benchmarks are added.

How are category weights determined?▾

Category weights reflect real-world demand: Coding and Reasoning each get 20%, General and Math each get 15%, Language and Multimodal each get 10%, and Safety and Agent each get 5%. Weights are renormalized across available categories so models are not penalized for missing data.

Can I compare specific models side by side?

AI Model Benchmarks

AI Model Benchmarks

Highlights

Intelligence

Context Window

Price

Leaderboard

Intelligence Tradeoffs

Intelligence vs. Output Cost ($/1M tokens)

Category Leaders

General Knowledge

Coding

Mathematics

Reasoning

Language

Multimodal

Safety

Agentic

Top Models per Task

Best in GPQA

Best in SWE-bench

Best in MATH

Best in MMLU-Pro

Best in HumanEval

Best in Chatbot Arena Elo

Best Model For Your Use Case

Coding & Engineering

Research & Analysis

Content & Copywriting

Best Value

Methodology

How it works

Coverage & confidence

Data sources

Frequently Asked Questions

Intelligence vs. Context Window

#	Model											Coverage
2—	Claude Fable 5 Anthropic🇺🇸1mo	97.7	98.1	—	97.2	—	—	—	$50	—	—	low
3—	Claude Mythos Preview Anthropic🇺🇸	92.1	97.7	—	94.6	83.3	—	91.8	—	—	—	medium
4—	Claude Opus 4.8 Anthropic🇺🇸1mo	91.9	93.2	—	92.5	—	—	91.9	$25	—	—	high
5—	GPT-5.6 Sol OpenAI🇺🇸1w	90.8	82.5	98.2	91.9	97.5	—	86.6	$30	—	—	medium
6—	M Kimi K3 Moonshot AI🇨🇳5dOSS	90.6	—	—	91.8	—	—	87.0	$15	—	—	high
9—	GPT-5.6 Terra OpenAI🇺🇸1w	88.1	77.6	97.4	90.5	97.0	—	82.6	$15	—	—	medium
10—	X Grok 4.5 xAI🇺🇸5d	86.7	82.8	—	90.6	—	—	—	$6	—	—	low
11—	Claude Sonnet 5 Anthropic🇺🇸3w	86.6	82.5	—	93.5	—	—	85.6	$15	—	—	medium
12—	Claude Opus 4.7 Anthropic🇺🇸3mo	86.0	86.1	—	91.5	80.1	—	89.3	$25	—	—	high
13—	X Grok-4 Heavy xAI🇺🇸	85.6	—	84.0	86.7	—	—	—	—	—	—	low
14—	GPT-5.1 High OpenAI🇺🇸8mo	84.9	—	83.5	85.8	—	—	—	—	—	—	low
15—	M Muse Spark 1.1 Meta🇺🇸1w	84.3	68.5	—	96.3	—	—	85.7	$4	—	—	medium
16—	GPT-5.1 Medium OpenAI🇺🇸8mo	82.1	—	82.1	—	—	—	—	$10	—	—	low
17—	B Seed 2.1 Pro ByteDance🇨🇳3w	82	71.0	—	92.2	—	—	84.3	—	—	—	high
18—	Claude Opus 4.6 Anthropic🇺🇸5mo	81.6	82.4	83.8	87	84.8	—	69.1	$25	—	—	high
19—	GPT-5 High OpenAI🇺🇸11mo	81.5	—	77.0	84.9	—	—	—	—	—	—	low
21—	ChatGPT-4o Latest OpenAI🇺🇸2y	80.7	—	—	80.7	—	—	—	—	—	—	low
22—	B Seed 2.0 Pro ByteDance🇨🇳5mo	80.5	74.9	82.0	86.7	—	—	—	$3	—	—	medium
23—	GPT-5.1 Codex High OpenAI🇺🇸8mo	79.9	—	79.9	—	—	—	—	—	—	—	low
24—	GPT-5.5 Pro OpenAI🇺🇸2mo	79.5	—	57.5	93.4	—	—	—	—	—	—	low
25—	Claude Sonnet 4.5 Anthropic🇺🇸9mo	79.1	92.8	64.7	79.9	72.5	—	—	$15	—	—	medium
26—	B Seed 2.1 Turbo ByteDance🇨🇳3w	79.0	68.3	—	91.3	—	—	81.2	—	—	—	high
28—	Z GLM-5.2 Zhipu AI🇨🇳1moOSS	78.5	71.6	—	90.2	—	—	—	$3	—	—	medium
29—	X Grok-3 xAI🇺🇸1y	78.4	—	75.0	81.5	—	—	77.0	$15	—	—	low
30—	GPT-5 Medium OpenAI🇺🇸11mo	78.2	—	68.0	85.8	—	—	—	—	—	—	low
31—	X MiMo-V2-Pro Xiaomi🇨🇳4mo	77.7	77.7	—	—	—	—	—	—	—	—	low
32—	Q Qwen3.7 Max Alibaba Cloud / Qwen Team🇨🇳2mo	77.6	76.5	—	81.6	76.5	—	—	$4	—	—	medium
33—	Z GLM-5V-Turbo Zhipu AI🇨🇳3mo	77.3	—	—	—	—	—	—	—	—	—	low
34—	GPT-5.6 Luna OpenAI🇺🇸1w	77.3	74.5	95.5	90.0	46.2	—	77.9	$6	—	—	medium
35—	ERNIE 5.0 Baidu🇨🇳6mo	77.2	—	64.7	75.3	92.2	—	—	—	—	—	medium
36—	GPT-5.2 Pro OpenAI🇺🇸7mo	77.1	—	84.0	72.7	—	—	—	—	—	—	medium
37—	X Grok-3 Mini xAI🇺🇸1y	76.6	—	71.1	80.7	—	—	—	—	—	—	low
38—	GPT-5.1 OpenAI🇺🇸8mo	76.3	74.5	57.0	85.8	—	—	89.6	$10	—	—	medium
39—	GPT-5.1 Instant OpenAI🇺🇸8mo	76.3	74.5	57.0	85.8	—	—	89.6	$10	—	—	medium
40—	GPT-5.1 Thinking OpenAI🇺🇸8mo	76.3	74.5	57.0	85.8	—	—	89.6	—	—	—	medium
41—	Claude Sonnet 4.6 Anthropic🇺🇸5mo	76.2	80.4	—	80.8	73.2	—	71.3	$15	—	—	high
42—	o1 OpenAI🇺🇸1y	75.7	49.0	94.2	80.3	86.7	—	—	$60	—	—	high
43—	Claude Opus 4 Anthropic🇺🇸1y	75.4	58.6	75.5	80.3	96.0	—	68.1	$75	—	—	high
44—	B Seed 2.0 Lite ByteDance🇨🇳5mo	75.3	68.9	74.6	82.2	—	—	—	—	—	—	low
45—	M Kimi K2-Thinking-0905 Moonshot AI🇨🇳10moOSS	74.8	66.6	84.0	84.5	—	—	—	—	—	—	medium
46—	M Kimi K2.6 Moonshot AI🇨🇳3moOSS	73.4	71.2	—	75.8	—	—	82.2	$4	—	—	high
47—	T Hy3 Tencent🇨🇳2wOSS	73.3	63.0	—	88.3	—	—	—	—	—	—	medium
48—	GPT-5.2 OpenAI🇺🇸7mo	73.3	81.1	71.3	70.3	74.2	—	83.1	$14	—	—	high
49—	Gemini 3.1 Pro Google🇺🇸5mo	73.3	68.4	—	89.6	53.5	—	82.2	$15	—	—	high
50—	GPT-5.5 OpenAI🇺🇸2mo	73.2	52.4	51.0	91.3	89.3	—	86.9	$30	—	—	high
51—	DeepSeek-V4-Pro-Max DeepSeek🇨🇳2moOSS	73.2	58.2	—	86.0	77.7	—	—	$3	—	—	high
52—	Z GLM-5 Zhipu AI🇨🇳5moOSS	73.2	77.3	—	—	—	—	—	$3	—	—	low
53—	o1-pro OpenAI🇺🇸1y	73.1	—	—	73.1	—	—	—	—	—	—	low
55—	S Step-3.5-Flash StepFun🇨🇳5moOSS	73.0	70.8	80.7	—	—	—	—	$0.40	—	—	low
56—	Gemini 3 Pro Google🇺🇸8mo	72.8	74.3	84.0	66.5	65.1	—	77.2	—	—	—	high
57—	GPT OSS 20B High OpenAI🇺🇸11moOSS	72.3	—	82.5	64.7	—	—	—	—	—	—	low
58—	Claude Opus 4.5 Anthropic🇺🇸7mo	72.0	82.5	—	61.5	78.0	—	—	—	—	—	medium
59—	Q Qwen3.7-Plus Alibaba Cloud / Qwen Team🇨🇳1mo	71.7	66.8	—	73.8	72.2	—	82.7	$1	—	—	high
60—	X MiMo-V2-Omni Xiaomi🇨🇳4mo	71.6	71.6	—	—	—	—	—	—	—	—	low
61—	M Kimi K2.5 Moonshot AI🇨🇳5moOSS	71.3	51.8	79.1	86.0	—	—	70.6	—	—	—	high
62—	GPT-5 Codex OpenAI🇺🇸10mo	71.0	71.0	—	—	—	—	—	—	—	—	low
63—	GPT-5.5 Instant OpenAI🇺🇸2mo	70.9	—	54.0	82.8	—	—	72.6	$30	—	—	medium
64—	Claude Opus 4.1 Anthropic🇺🇸11mo	70.5	75.2	47.9	76.2	73.8	—	—	—	—	—	medium
65—	S Step3-VL-10B StepFun🇨🇳6moOSS	70.4	—	65.9	—	—	—	77.2	—	—	—	low
66—	Q Qwen3.5-122B-A10B Alibaba Cloud / Qwen Team🇨🇳4moOSS	70.3	65.6	—	83.6	63.7	—	73.8	—	—	—	high
67—	Gemini 3 Flash Google🇺🇸7mo	70.3	77.7	83.7	66.0	62.6	—	73.7	$3	—	—	high
69—	Q Qwen3.5-27B Alibaba Cloud / Qwen Team🇨🇳4moOSS	69.4	66.5	—	83.6	60.6	—	73.4	$2	—	—	high
70—	GPT-5.1 Codex OpenAI🇺🇸8mo	69.3	69.3	—	—	—	—	—	—	—	—	low
71—	Gemma 4 31B Google🇺🇸3moOSS	69.2	—	—	61.2	76.3	—	74.5	$0.38	—	—	medium
72—	MiniMax M3 MiniMax🇨🇳1moOSS	69.1	68.3	—	—	—	—	77.2	$1	—	—	medium
73—	Claude Sonnet 4 Anthropic🇺🇸1y	68.9	54.5	66.8	71.6	90.0	—	63.7	$15	—	—	high
74—	X Grok 4 Fast xAI🇺🇸10mo	68.5	—	73.0	55.4	98.6	—	—	—	—	—	medium
75—	GPT-5.4 OpenAI🇺🇸4mo	68.3	47.2	69.1	82.2	—	—	83.5	$15	—	—	high
76—	GPT OSS 120B High OpenAI🇺🇸11moOSS	68.3	—	73.8	76.2	52.1	—	—	—	—	—	low
77—	M Kimi K2.7 Code Moonshot AI🇨🇳1moOSS	68.1	—	—	—	—	—	—	$4	—	—	low
78—	DeepSeek-V3.2 DeepSeek🇨🇳7moOSS	67.9	68.0	74.8	75.2	—	—	—	—	—	—	medium
79—	MAI-Thinking-1 Microsoft🇺🇸1mo	67.8	45.3	80.3	81.0	—	—	—	—	—	—	medium
80—	Q Qwen3.5-397B-A17B Alibaba Cloud / Qwen Team🇨🇳5moOSS	67.7	74.7	—	66.1	70.4	—	—	—	—	—	medium
81—	M Kimi K2 0905 Moonshot AI🇨🇳10mo	67.6	—	—	67.6	—	—	—	—	—	—	low
82—	Z GLM-4.7 Zhipu AI🇨🇳7moOSS	67.3	57.0	78.5	79.3	—	—	—	—	—	—	medium
83—	L K-EXAONE-236B-A23B LG AI Research🇰🇷6mo	67.0	—	74.3	—	59.8	—	—	—	—	—	low
84—	Nova 2 Pro Amazon🇺🇸7mo	67.0	67.3	73.5	76.9	—	—	37.0	—	—	—	medium
85—	Z GLM-5.1 Zhipu AI🇨🇳3moOSS	66.9	51.3	—	86.3	—	—	—	$4	—	—	medium
86—	M Muse Spark Meta🇺🇸3mo	66.8	48.3	—	75.8	—	—	85.9	—	—	—	high
87—	Gemini 2.0 Flash Thinking Google🇺🇸1y	66.8	—	—	64.7	—	—	71.1	—	—	—	low
88—	Q Qwen3-Next-80B-A3B-Thinking Alibaba Cloud / Qwen Team🇨🇳10moOSS	66.8	—	66.1	70.1	—	—	—	—	—	—	low
89—	Llama 3.1 405B Meta🇺🇸1yOSS	66.3	69.3	72.0	48.9	67.9	87.7	—	$3	—	—	high
90—	Q Qwen3.6 Plus Alibaba Cloud / Qwen Team🇨🇳3mo	66.1	60.0	—	67.3	73.8	—	76.8	$3	—	—	high
91—	Q Qwen3.5-35B-A3B Alibaba Cloud / Qwen Team🇨🇳4moOSS	65.9	59.3	—	82.0	57.8	—	70.8	—	—	—	high
92—	Mistral Medium 3.5 Mistral AI🇫🇷2moOSS	64.9	77.0	63.4	—	—	—	—	$8	—	—	low
94—	Q Qwen3-Coder 480B A35B Instruct Alibaba Cloud / Qwen Team🇨🇳1yOSS	63.6	60.2	—	—	—	—	—	—	—	—	low
95—	GPT-5 OpenAI🇺🇸11mo	63.5	71.8	57.1	60.3	—	—	79.2	—	—	—	high
96—	DeepSeek-V3.2-Exp DeepSeek🇨🇳9moOSS	63.5	58.6	68.6	51.0	98.8	—	—	—	—	—	medium
97—	X Grok-4 xAI🇺🇸1y	63.2	—	72.6	56.2	—	—	—	—	—	—	medium
98—	M LongCat-Flash-Thinking-2601 Meituan🇨🇳6moOSS	63.1	61.1	83.5	57.0	—	—	—	—	—	—	medium
99—	DeepSeek R1 Zero DeepSeek🇨🇳1yOSS	63.0	—	—	63.0	—	—	—	—	—	—	low
100—	X Grok Code Fast 1 xAI🇺🇸10mo	63.0	63.0	—	—	—	—	—	$2	—	—	low
101—	DeepSeek-V3.2 (Thinking) DeepSeek🇨🇳7moOSS	62.3	68.0	74.8	58.3	—	—	—	—	—	—	medium
102—	Q Qwen3.5-9B Alibaba Cloud / Qwen Team🇨🇳4moOSS	62.0	—	—	77.4	41.5	—	—	—	—	—	low
103—	MiniMax M2.5 MiniMax🇨🇳5moOSS	60.6	57.8	—	—	—	—	—	$1	—	—	low
105—	M LongCat-Flash-Thinking Meituan🇨🇳10moOSS	60.5	36.3	70.8	77.1	—	—	—	—	—	—	low
106—	Claude 3.5 Sonnet Anthropic🇺🇸2y	60.4	48.4	58.6	68.4	80.5	31.6	52.2	$15	—	—	high
107—	Q Qwen3-235B-A22B-Instruct-2507 Alibaba Cloud / Qwen Team🇨🇳12moOSS	60.4	—	33.7	70.6	73.3	—	—	—	—	—	low
108—	Gemini 3.1 Flash-Lite Google🇺🇸4mo	60.1	—	—	52.6	68.2	—	63.0	$2	—	—	medium
109—	DeepSeek-V3.2-Speciale DeepSeek🇨🇳7moOSS	60.0	68.0	79.0	50.4	—	—	—	—	—	—	medium