Updated April 2026

AI LLM Leaderboard 2026 — Model Rankings & Scores

The TC Score ranks 16 leading large language models across five weighted capability dimensions - giving you a single, comparable number grounded in public benchmarks, community evaluations, and our own assessments as of April 2026.

Benchmark subjectivity notice: Rankings are necessarily subjective. Different tasks, prompt styles, and evaluation harnesses can shift results meaningfully. Always run your own evals for mission-critical use cases. Scores are approximate composites, not official vendor claims.

How the TC Score Is Calculated

The TC Score is a weighted composite of five capability dimensions. Each dimension is scored 0–100 and combined using the weights below.

Reasoning 25%
Coding 25%
Writing 20%
Multimodal 15%
Value 15%

Top 3 - April 2026 Podium

2
Silver
GPT-5.4
OpenAI
96.2
  • Best-in-class coding
  • Excellent tool use
  • Strong ecosystem
1
Gold
Claude Opus 4.6
Anthropic
97.8
  • Best reasoning + writing
  • Top multimodal capability
  • 200K context window
3
Bronze
Gemini 3.1 Ultra
Google
95.1
  • 2M token context
  • Excellent multimodal
  • Native Google Search

Full Rankings - All 16 Models

Rank Model Provider Context Category Scores TC Score Trend
🥇 Claude Opus 4.6 Best Overall Anthropic 200K
R
98
C
97
W
96
M
95
V
100
97.8 ⇧
🥈 GPT-5.4 Best Coding OpenAI 128K
R
97
C
98
W
95
M
94
V
96
96.2 ⇧
🥉 Gemini 3.1 Ultra Best Context Google 2M
R
96
C
95
W
94
M
98
V
92
95.1 ⇧
#4 o3-pro Best Reasoning OpenAI 200K
R
99
C
96
W
88
M
85
V
82
93.4 ⇧
#5 Claude Sonnet 4.6 Best Value Anthropic 200K
R
92
C
94
W
93
M
90
V
97
91.2 —
#6 Gemini 3.1 Pro Google 1M
R
90
C
91
W
92
M
94
V
90
90.8 ⇧
#7 GPT-5.4 mini Speed Pick OpenAI 128K
R
86
C
89
W
88
M
84
V
95
87.6 ⇧
#8 Grok 3.5 xAI 128K
R
85
C
87
W
86
M
82
V
86
85.3 ⇧
#9 Llama 4 Maverick Open Source Meta 128K
R
83
C
86
W
84
M
88
V
92
84.6 ⇧
#10 DeepSeek V4 Best Price DeepSeek 64K
R
84
C
87
W
80
M
75
V
98
84.0 —
#11 Claude Haiku 4.5 Anthropic 200K
R
78
C
80
W
82
M
78
V
99
79.8 —
#12 Gemini 3.1 Flash Google 1M
R
76
C
78
W
80
M
82
V
98
78.2 —
#13 Qwen3 72B Alibaba 128K
R
80
C
83
W
78
M
72
V
95
77.4 ⇧
#14 Mistral Large 2 Mistral 128K
R
75
C
78
W
76
M
70
V
90
74.8 —
#15 Phi-4 Microsoft 16K
R
72
C
76
W
70
M
65
V
96
71.4 —
#16 Llama 3.3 70B Meta 128K
R
70
C
73
W
72
M
68
V
97
70.2 ⇩

Category Leaders

Reasoning Leaders

1
o3-pro
99
2
Claude Opus 4.6
98
3
GPT-5.4
97
4
Gemini 3.1 Ultra
96
5
Claude Sonnet 4.6
92

Coding Leaders

1
GPT-5.4
98
2
Claude Opus 4.6
97
3
o3-pro
96
4
Gemini 3.1 Ultra
95
5
Claude Sonnet 4.6
94

Writing Leaders

1
Claude Opus 4.6
96
2
GPT-5.4
95
3
Gemini 3.1 Ultra
94
4
Claude Sonnet 4.6
93
5
Gemini 3.1 Pro
92

Multimodal Leaders

1
Gemini 3.1 Ultra
98
2
Claude Opus 4.6
95
3
GPT-5.4
94
4
Gemini 3.1 Pro
94
5
Claude Sonnet 4.6
90

Speed & Value

1
Claude Opus 4.6
100
2
Claude Haiku 4.5
99
3
DeepSeek V4
98
4
Gemini 3.1 Flash
98
5
Claude Sonnet 4.6
97