What is preference Elo on CompareLLM?

Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent.

How often does CompareLLM update?

Once a day at 06:00 UTC. A Vercel Hobby cron hits /api/cron/ingest, pulls OpenRouter, LiveBench, official SWE-bench JSON, and Arena Elo, then refreshes pages if a number moved. Humans do not retype cells.

Do you run SWE-bench yourselves?

No. V1 is an aggregator plus optional first-party latency pings. Missing cells stay blank. We do not scrape Artificial Analysis and we do not invent a composite intelligence index.

How do new compare pages appear?

Automatically. Two indexable models that share at least three metrics get /compare/{a}-vs-{b}. Thin pairs still render but are noindex so we do not build doorway pages.

Is public login the same as admin?

No. Anyone can register to vote, comment, and save models. Publishing /news, running ingest by hand, promoting preview models, and attaching aliases is an allowlisted admin desk.

Updated Aug 17, 2026·Live Ingest

Compare AI Models on Verified Benchmarks, Speed & Real Cost

Name: CompareLLM AI model leaderboard
Creator: CompareLLM
License: https://creativecommons.org/licenses/by/4.0/

Independent crowd preference Elo, SWE-bench coding tests, and live $/1M API pricing with zero synthetic bias.

Best coding Best cheap

Intelligence vs API Price Tradeoff

Quality vs. Price Pareto Frontier

Top-Left = Best Value

Plotted by intelligence (Elo) vs cost (Out $). The teal line connects models with unbeatable price-to-performance.

Leaderboard TableTable↓

↑ Up: Crowd vote (Elo)← Left: Cheaper

Vertical ↑

Crowd vote (Elo)

Lost more votes ↓People liked it more ↑

Two hidden answers. A person picks the one they like. Higher Elo means more wins — not a school test or a coding exam.

Horizontal →

Output (answer) price / 1 million tokens (USD)

← CheaperMore expensive →

Cost to generate the answer. Provider list prices are USD

Best-Value Frontier Line (9 Models)Models defining the teal efficiency boundary

Brand:

💬 LLMs 🎨 Text-to-Image ModelsText-to-ImageNew

Sortable LLM Leaderboard

Live AI Model Leaderboard

Top 10 of 26 models.Top 26 of 26 models. Sortable rankings from dated verified benchmarks.

Compare every row vs:

All Models Frontier Open Weights 🇨🇳 China

1GLM-5.3🇨🇳 CNFrontier

Zhipu

Elo1,558

SWE76.4%

In /1M$0.5

Out /1M$1.8

vs Claude Opus 5 Specs

2DeepSeek V4 Pro🇨🇳 CNOpenFrontier

DeepSeek

Elo1,536

SWE71.6%

In /1M$1.32

Out /1M$3.96

vs Claude Opus 5 Specs

3Kimi K3🇨🇳 CNFrontier

Moonshot

Elo1,528

SWE73.8%

In /1M$3

Out /1M$15

vs Claude Opus 5 Specs

4Qwen 3 Max🇨🇳 CNFrontier

Alibaba

Elo1,518

SWE67.8%

In /1M$0.78

Out /1M$3.9

vs Claude Opus 5 Specs

5MoonshotAI: Kimi K2.5🇨🇳 CN

Moonshot

Elo1,515

SWE71.3%

In /1M$0.57

Out /1M$2.85

vs Claude Opus 5 Specs

6Qwen QwQ 32B🇨🇳 CNOpenFrontier

Alibaba

Elo1,495

SWE67.5%

In /1M$0.3

Out /1M$1.2

vs Claude Opus 5 Specs

7GLM-5.2🇨🇳 CNOpenFrontier

Zhipu

Elo1,492

SWE69.3%

In /1M$1.19

Out /1M$3.74

vs Claude Opus 5 Specs

8DeepSeek V4 Flash🇨🇳 CNOpenFrontier

DeepSeek

Elo1,490

SWE71.2%

In /1M$0.14

Out /1M$0.28

vs Claude Opus 5 Specs

9MiniMax M2.5🇨🇳 CNOpenFrontier

MiniMax

Elo1,478

SWE75.8%

In /1M$0.22

Out /1M$0.9

vs Claude Opus 5 Specs

10Yi-Lightning🇨🇳 CNFrontier

01.AI

Elo1,475

SWE63.4%

In /1M$0.14

Out /1M$0.14

vs Claude Opus 5 Specs

#	Model	Elo	LiveBench	SWE-bench	tok/s	In $/1M	Out $/1M
1	GLM-5.3🇨🇳 CNFrontier Zhipu	1,558	70.8%	76.4%	90 tok/s	$0.50 / 1M	$1.80 / 1M
2	DeepSeek V4 Pro🇨🇳 CNOpenFrontier DeepSeek	1,536	70.1%	71.6%	70 tok/s	$1.32 / 1M	$3.96 / 1M
3	Kimi K3🇨🇳 CNFrontier Moonshot	1,528	69.5%	73.8%	68 tok/s	$3.00 / 1M	$15.00 / 1M
4	Qwen 3 Max🇨🇳 CNFrontier Alibaba	1,518	69.4%	67.8%	80 tok/s	$0.78 / 1M	$3.90 / 1M
5	MoonshotAI: Kimi K2.5🇨🇳 CN Moonshot	1,515	67.2%	71.3%	72 tok/s	$0.57 / 1M	$2.85 / 1M
6	Qwen QwQ 32B🇨🇳 CNOpenFrontier Alibaba	1,495	68%	67.5%	68 tok/s	$0.30 / 1M	$1.20 / 1M
7	GLM-5.2🇨🇳 CNOpenFrontier Zhipu	1,492	65.2%	69.3%	84 tok/s	$1.19 / 1M	$3.74 / 1M
8	DeepSeek V4 Flash🇨🇳 CNOpenFrontier DeepSeek	1,490	67.8%	71.2%	142 tok/s	$0.14 / 1M	$0.28 / 1M
9	MiniMax M2.5🇨🇳 CNOpenFrontier MiniMax	1,478	64.4%	75.8%	90 tok/s	$0.22 / 1M	$0.90 / 1M
10	Yi-Lightning🇨🇳 CNFrontier 01.AI	1,475	65%	63.4%	130 tok/s	$0.14 / 1M	$0.14 / 1M
11	Z.ai: GLM 5V Turbo🇨🇳 CN Zhipu	1,475	63.8%	65.2%	112 tok/s	$1.20 / 1M	$4.00 / 1M
12	Seed 2.1 Turbo🇨🇳 CN ByteDance	1,472	64.8%	63.2%	160 tok/s	$0.50 / 1M	$2.50 / 1M
13	Doubao Pro 1.5🇨🇳 CNFrontier ByteDance	1,470	64%	62.1%	105 tok/s	$0.11 / 1M	$0.28 / 1M
14	MoonshotAI: Kimi K2 0711🇨🇳 CN Moonshot	1,465	62.8%	65.8%	85 tok/s	$0.57 / 1M	$2.30 / 1M
15	Ernie 4.5 Turbo🇨🇳 CN Baidu	1,450	61%	57%	95 tok/s	$0.30 / 1M	$1.20 / 1M
16	Qwen 2.5 Plus🇨🇳 CN Alibaba	1,445	63.5%	59.2%	110 tok/s	$0.25 / 1M	$1.00 / 1M
17	Qwen3 235B🇨🇳 CNOpen Alibaba	1,440	63.5%	59.4%	74 tok/s	$0.45 / 1M	$1.82 / 1M
18	Yi-Large🇨🇳 CN 01.AI	1,430	60.5%	53.8%	90 tok/s	$0.30 / 1M	$0.30 / 1M
19	Qwen 2.5 Coder 32B🇨🇳 CNOpen Alibaba	1,425	64.8%	65.2%	110 tok/s	$0.18 / 1M	$0.72 / 1M
20	Qwen2.5 Coder 32B Instruct🇨🇳 CN Alibaba	1,425	64.8%	65.2%	—	$0.66 / 1M	$1.00 / 1M
21	Kimi Chat 1.5🇨🇳 CN Moonshot	1,420	58%	51%	95 tok/s	$0.20 / 1M	$0.80 / 1M
22	Qwen3.8 27B🇨🇳 CNOpen Alibaba	1,398	60.6%	58.8%	152 tok/s	$0.45 / 1M	$3.20 / 1M
23	DeepSeek Coder V2🇨🇳 CNOpen DeepSeek	1,365	58.6%	60.5%	75 tok/s	$0.14 / 1M	$0.28 / 1M
24	Doubao Lite 1.5🇨🇳 CN ByteDance	1,360	54%	46%	180 tok/s	$0.04 / 1M	$0.10 / 1M
25	DeepSeek R1🇨🇳 CNOpen DeepSeek	1,358	59%	65.2%	62 tok/s	$0.70 / 1M	$2.50 / 1M
26	DeepSeek V3🇨🇳 CNOpen DeepSeek	1,310	55.4%	48.6%	66 tok/s	$0.26 / 1M	$1.03 / 1M

26 models in this filter · sorted by elo

All benchmark leaderboards Intent rankings

US vs China AI FrontierGap Closed & Surpassing on Cost

How Close is China's AI to US Flagships?

Comparing 🇺🇸 Claude Opus 5 (Anthropic) vs 🇨🇳 GLM-5.3 (Zhipu). Quality is virtually tied, while China is surpassing on price and open weights.

Showdown: Claude vs GLM-5.3

General IntelligenceChatbot Arena human preference score

🇺🇸1624 Elo

🇨🇳1558 Elo

Near Parity (99%)

Software EngineeringSWE-bench real GitHub bug resolution

🇺🇸79.2%

🇨🇳76.4%

Near Parity (97%)

API Pricing & CostOutput price per 1 Million tokens

🇺🇸$50.00

🇨🇳$1.80

Surpassing (27.8× cheaper)

Model AvailabilityWeights download & self-hosting

🇺🇸Closed API

🇨🇳Open Weights

Surpassing (Open License)

Takeaway:GLM-5.3 gives you 96% of Claude Opus 5's capability at 96% lower cost.

See full benchmark test logs

Leaders Matchup

Open Claude Opus 5 vs GPT-5.6 Sol fact sheetOpen showdown fact sheet|View all comparisons

Capability Matchup Radar

Multi-Dimensional Capability Radar

Each spoke is a skill. Farther from the center is better (0–100th percentile). Tap any dot to inspect details.

Tap any node to inspect

Percentile 0–100

Dimensional AdvantageEven 3–3

Verified Intelligence & Dispatches

Recent AI News & Benchmark Briefings

Independent analysis on model promotions, price reductions, and verified score movements.

View all dispatches (51)

GLM-5.2 vs GLM 5V Turbo Price and SWE-bench 2026

news

AIEval EditorialAug 16, 2026

GLM-5.2 vs GLM 5V Turbo Price and SWE-bench 2026

GLM-5.2 undercuts GLM 5V Turbo by 74.2% on input at $0.31/1M, with 1,492 Elo and 69.3% SWE-bench in dated 2026 snapshots.

Shipped & Updated

New and Refreshed Models

View all 460 catalog models

Zhipu2026-08-14

GLM-5.3

Z.ai Aug 14 2026 post-train of the GLM-5.2 744B base. Coding-plan live; open weights promised after a two-week safety review (z.ai/blog/glm-5.3).

Alibaba2026-08-14

Qwen3.8 27B

Alibaba open-weight 27B drop dated Aug 14 2026. Dense enough to self-host; not a frontier MoE.

How to read this leaderboard

Plain-English methodology and leaderboard answers

: Preference Elo is a crowd vote from LMArena / Arena. People see two hidden answers and pick the one they like more. The model that wins more often gets a higher Elo. That means people preferred it — not that it passed a school test. It is not SWE-bench, not accuracy, and not a number we invent.

Compare AI Models on Verified Benchmarks, Speed & Real Cost

Quality vs. Price Pareto Frontier

Live AI Model Leaderboard

How Close is China's AI to US Flagships?

Leaders Matchup

Multi-Dimensional Capability Radar

Recent AI News & Benchmark Briefings

GLM-5.2 vs GLM 5V Turbo Price and SWE-bench 2026

New and Refreshed Models

How to read this leaderboard

Head-to-Head Comparison Studio

Claude Opus 5 vs GPT-5.6 Sol

Claude Sonnet 5 vs DeepSeek V4

Qwen3.8 vs Llama 4 Scout

GLM-5.3 vs Gemini 3.7 Flash

Haiku 4.5 vs Llama 4 Scout: cheap Claude versus cheap open Llama

Qwen3.8 27B (Aug 14): dense enough to self-host, not a frontier MoE