LLM Leaderboard 2025

The definitive ranking of AI models including Gemini 3, GPT-5, and Grok. Comparison based on coding, reasoning, context, and speed.

Model
Developer
Context
Release
Multimodality
Code Gen
Reasoning
Speed
Rank
Gemini 3 Pro
Google / DeepMind
~1M / 64K out
2025
Full Multimodal
Excellent
State-of-the-art
Med-Fast
1
grok-4.1-thinking
xAI
~128K
2025
Text, Images
Excellent
Excellent
Very Fast
2
grok-4.1
xAI
~128K
2025
Text, Images
Excellent
Excellent
Very Fast
3
GPT-5 (high)
OpenAI
400K
2025
Text & Vision
Excellent
Very Strong
Very Fast
4
Gemini 2.5 Pro
Google
~1M
2025
Full Multimodal
Excellent
Excellent
Fast
5
Claude 3.5 Sonnet (Think)
Anthropic
~32K
2025
Text, Images
Excellent
Excellent
Medium
6
Claude Opus 4.1 (Think)
Anthropic
~32K
2025
Text, Images
Excellent
Excellent
Medium
7
Claude 3.5 Sonnet
Anthropic
~32K
2025
Text, Images
Excellent
Excellent
Medium
8
GPT-4.5 (Preview)
OpenAI
~128K
2025
Text & Vision
Excellent
Very Good
Fast
9
Claude Opus 4.1
Anthropic
~16K
2025
Text, Images
Excellent
Very Good
Medium
10
GPT-5.1 (Chat)
OpenAI
400K
2025
Text & Vision
Excellent
Excellent
Very Fast
11
Gemini 1.5 Flash
Google
~1M
2024
Full Multimodal
Good
Good
Very Fast
12
GPT-4 Turbo
OpenAI
~128K
2023
Text, Images
Excellent
Excellent
Fast
13
Llama 3 70B
Meta
~8K
2024
Primarily Text
Good
Good
Very Fast
14
Mistral Large
Mistral AI
32K
2024
Text
Excellent
Good
Fast
15
Qwen3 / Qwen2
Alibaba
128K
2025
Text, Images
Excellent
Good
Fast
16
DBRX Instruct
Databricks
~32K
2024
Text
Excellent
Good
Fast
17
Phi-3
Microsoft
~128K
2024
Text
Good
Avg-Good
Very Fast
18
Gemma 7B
Google
~8K
2024
Text
Good
Good
Very Fast
19
Jurassic-2 Ultra
AI21 Labs
~8K
2023
Text
Average
Average
Fast
20