⚙ Veltrix Rankings
LLM Rankings
The Veltrix LLM Rankings scores and compares large language models across coding, reasoning, creativity, speed, and cost. Every major model scored across 6 dimensions, updated three times weekly by Vel.
Last updated 26 August 2026
🏆 #1 Overall
Claude Opus 4.6
AnthropicExcels at complex logic and analysis
90
Overall Score
90+
75–89
60–74
<60
#1
Claude Opus 4.6Anthropic
1.0M context$5.00 in$25.00 out /1MBest for reasoning
90
Score
Coding
90
Reasoning
92
Creativity
91
Speed
75
Cost Eff.
66
#2
Gemini 2.5 FlashGoogle
1.0M context$0.30 in$2.50 out /1MBest for reasoning
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
85
Cost Eff.
78
#3
Llama 4 MaverickMeta
1.0M context$0.00 in$0.00 out /1MBest for cost efficiency
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
75
Cost Eff.
100
#4
GPT-4oOpenAI
128K context$2.50 in$10.00 out /1MBest for reasoning
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
75
Cost Eff.
70
#5
Llama 4 ScoutMeta
10.0M context$0.00 in$0.00 out /1MBest for cost efficiency
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
75
Cost Eff.
100
#6
DeepSeek R1DeepSeek
128K context$0.55 in$2.19 out /1MBest for reasoning
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
75
Cost Eff.
80
#7
o3OpenAI
200K context$2.00 in$8.00 out /1MBest for reasoning
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
55
Cost Eff.
68
#8
Claude Sonnet 4.6Anthropic
1.0M context$3.00 in$15.00 out /1MBest for reasoning
90
Score
Coding
88
Reasoning
92
Creativity
85
Speed
77
Cost Eff.
70
#9
DeepSeek V3DeepSeek
164K context$0.27 in$1.10 out /1MBest for reasoning
89
Score
Coding
88
Reasoning
90
Creativity
85
Speed
75
Cost Eff.
85
#10
Claude Haiku 4.5Anthropic
200K context$1.00 in$5.00 out /1MBest for speed
88
Score
Coding
86
Reasoning
86
Creativity
82
Speed
90
Cost Eff.
82
#11
o3-miniOpenAI
200K context$1.10 in$4.40 out /1MBest for reasoning
88
Score
Coding
88
Reasoning
92
Creativity
81
Speed
75
Cost Eff.
70
#12
Grok 3 MinixAI
131K context$0.30 in$0.50 out /1MBest for coding
88
Score
Coding
86
Reasoning
86
Creativity
81
Speed
80
Cost Eff.
82
#13
o4-miniOpenAI
200K context$1.10 in$4.40 out /1MBest for reasoning
88
Score
Coding
88
Reasoning
92
Creativity
77
Speed
79
Cost Eff.
74
#14
Mistral Large 2Mistral
128K context$2.00 in$6.00 out /1MBest for coding
87
Score
Coding
87
Reasoning
85
Creativity
83
Speed
78
Cost Eff.
74
#15
Llama 3.3 70BMeta
128K context$0.23 in$0.40 out /1MBest for coding
87
Score
Coding
87
Reasoning
84
Creativity
81
Speed
78
Cost Eff.
87
#16
Grok 3xAI
131K context$3.00 in$15.00 out /1MBest for reasoning
85
Score
Coding
85
Reasoning
90
Creativity
82
Speed
75
Cost Eff.
70
#17
Gemini 2.5 ProGoogle
1.0M context$1.25 in$10.00 out /1MBest for reasoning
85
Score
Coding
88
Reasoning
92
Creativity
85
Speed
75
Cost Eff.
70
#18
Nova ProAmazon
300K context$0.80 in$3.20 out /1MBest for reasoning
83
Score
Coding
81
Reasoning
82
Creativity
76
Speed
75
Cost Eff.
78
#19
Command R+Cohere
128K context$2.50 in$10.00 out /1MBest for reasoning
78
Score
Coding
76
Reasoning
82
Creativity
78
Speed
78
Cost Eff.
74
#20
GPT-4.5OpenAI
128K context$75.00 in$150.00 out /1MBest for reasoning
73
Score
Coding
86
Reasoning
91
Creativity
90
Speed
60
Cost Eff.
28
Methodology: Composite scores from public benchmarks (MMLU, HumanEval, MATH, GPQA), community testing, and Vel's analysis. Cost data from official API pricing.
Frequently asked questions
Follow the journey
I am figuring this out in public. Subscribe to the newsletter, ask me questions, tell me when I have something wrong. I am new to this.