Skytells
  • Home
  • Changelog
Skytells

Addressing the world's greatest challenges with AI. Enterprise research, foundation models, and infrastructure trusted by organizations worldwide since 2012.

Get Started

  • Console
  • Learn
  • Pricing
  • ModelsNew
  • EveNew
  • CLI

Platform

  • DropsNew
  • Cognition
  • Cloud AgentsNew
  • AI Solutions
  • Infrastructure
  • Edge Network

Resources

  • Documentation
  • Blog
  • Changelog
  • AI Leaderboard
  • Research
  • Status

Company

  • About
  • Careers
  • Trust Center
  • Legal
  • Privacy Policy

© 2012–2026 Skytells, Inc. All rights reserved.

Live rankings

AI Model Leaderboard

Every major AI model ranked across benchmark quality, inference speed, agentic capability, programming aptitude, and cost efficiency — updated continuously from published evaluation data.

Explore full leaderboardBrowse model catalog

418

Tracked models

39

Providers

330

Benchmarked

32.2

Avg. index

OverallBenchmarksInferenceAgenticProgrammingValue / Price

418 models

RankModelProviderScoreBenchmarksInferenceAgenticProgrammingValuePrice
261

Gemini 1.5 Pro

gemini-1.5-pro

multimodalvisionmulti-input reasoning
Google

24.5

overall

24.50.00.00.00.0N/A
262

QwQ-32B

qwq-32b

textinference
AAlibaba Cloud / Qwen Team

24.5

overall

24.50.00.00.00.0N/A
263

QwQ-32B-Preview

qwq-32b-preview

textinference
AAlibaba Cloud / Qwen Team

24.5

overall

24.50.00.00.00.0N/A
264

IBM Granite 4.2 3B

granite-4.2-3b

codeprogrammingtool use
IIBM

24.4

overall

23.014.90.00.097.3$0.03 in / $0.12 out
265

Gemma 3 4B

gemma-3-4b-it

multimodalvisionmulti-input reasoning
Google

24.4

overall

4.114.90.00.096.7$0.05 in / $0.1 out
266

Llama 3.1 8B Instruct

llama-3.1-8b-instruct

textinference
MMeta

24.2

overall

2.714.90.00.099.5$0.02 in / $0.04 out
267

DeepSeek-V3

deepseek-v3

codeprogrammingtool use
DeepSeek

24.1

overall

24.721.90.07.762.0$0.32 in / $0.89 out
268

Grok-2

grok-2

multimodalvisionmulti-input reasoning
xAI

24.1

overall

24.10.00.00.00.0N/A
269

GPT-4o

gpt-4o-2024-08-06

multimodalvisionmulti-input reasoning
OpenAI

23.9

overall

27.556.714.93.228.3
270

Qwen3 VL 32B Instruct

qwen3-vl-32b-instruct

multimodalvisionmulti-input reasoning
AAlibaba Cloud / Qwen Team

23.6

overall

24.20.023.00.00.0
271

Gemini 2.0 Flash-Lite

gemini-2.0-flash-lite

multimodalvisionmulti-input reasoning
Google

23.0

overall

23.00.00.00.00.0
272

Sarvam-105B

sarvam-105b

codeprogrammingtool use
SSarvam AI

22.8

overall

39.70.015.88.90.0N/A
273

DeepSeek R1 Distill Qwen 32B

deepseek-r1-distill-qwen-32b

textinference
DeepSeek

22.4

overall

22.40.00.00.00.0N/A
274

DeepSeek-R1-0528

deepseek-r1-0528

codeprogrammingtool use
DeepSeek

22.1

overall

46.321.90.04.945.1$0.5 in / $2.15 out
275

GLM-4.5-Air

glm-4.5-air

codeprogrammingtool use
ZZhipu AI

22.1

overall

23.50.024.218.00.0N/A
276

Llama 3.1 70B Instruct

llama-3.1-70b-instruct

textinference
MMeta

22.0

overall

9.314.90.00.068.8$0.4 in / $0.4 out
277

Nemotron Nano 9B v2

nvidia-nemotron-nano-9b-v2

textinference
NNVIDIA

21.8

overall

21.80.00.00.00.0N/A
278

Mistral Large 3 (675B Base)

mistral-large-3-675b-base-2512

multimodalvisionmulti-input reasoning
Mistral AI

21.7

overall

21.70.00.00.00.0
279

Mistral Large 3 (675B Instruct 2512 Eagle)

mistral-large-3-675B-instruct-2512-eagle

multimodalvisionmulti-input reasoning
Mistral AI

21.7

overall

21.70.00.00.00.0
280

Mistral Large 3 (675B Instruct 2512 NVFP4)

mistral-large-3-675b-instruct-2512-nvfp4

multimodalvisionmulti-input reasoning
Mistral AI

21.7

overall

21.70.00.00.00.0
261

Gemini 1.5 Pro

Google

24.5

N/A

262
A

QwQ-32B

Alibaba Cloud / Qwen Team

24.5

N/A

263
A

QwQ-32B-Preview

Alibaba Cloud / Qwen Team

24.5

N/A

264

Page 14 of 21 · 418 models

PreviousNext

Want benchmark charts, model comparison, and pricing analytics?

Sign in to access the full interactive leaderboard with deep benchmark breakdowns and model comparison tools.

Open full leaderboard

Rankings are based on multi-dimensional evaluation across benchmark quality, inference efficiency, and cost-per-output. Scores are updated continuously and may differ from individual third-party benchmarks.

$2.5 in / $10 out
N/A
N/A
N/A
N/A
N/A
I

IBM Granite 4.2 3B

IBM

24.4

$0.03 in / $0.12 out

265

Gemma 3 4B

Google

24.4

$0.05 in / $0.1 out

266
M

Llama 3.1 8B Instruct

Meta

24.2

$0.02 in / $0.04 out

267

DeepSeek-V3

DeepSeek

24.1

$0.32 in / $0.89 out

268

Grok-2

xAI

24.1

N/A

269

GPT-4o

OpenAI

23.9

$2.5 in / $10 out

270
A

Qwen3 VL 32B Instruct

Alibaba Cloud / Qwen Team

23.6

N/A

271

Gemini 2.0 Flash-Lite

Google

23.0

N/A

272
S

Sarvam-105B

Sarvam AI

22.8

N/A

273

DeepSeek R1 Distill Qwen 32B

DeepSeek

22.4

N/A

274

DeepSeek-R1-0528

DeepSeek

22.1

$0.5 in / $2.15 out

275
Z

GLM-4.5-Air

Zhipu AI

22.1

N/A

276
M

Llama 3.1 70B Instruct

Meta

22.0

$0.4 in / $0.4 out

277
N

Nemotron Nano 9B v2

NVIDIA

21.8

N/A

278

Mistral Large 3 (675B Base)

Mistral AI

21.7

N/A

279

Mistral Large 3 (675B Instruct 2512 Eagle)

Mistral AI

21.7

N/A

280

Mistral Large 3 (675B Instruct 2512 NVFP4)

Mistral AI

21.7

N/A