Skip to main content

Best AI Models in 2026 — Overall Rankings

BenchLM.ai now distinguishes provisional overall ranking from verified overall ranking. The provisional score is a normalized weighted average across 8 benchmark categories: agentic (22%), coding (20%), reasoning (17%), knowledge (12%), multimodal & grounded (12%), multilingual (7%), instruction following (5%), and math (5%), using non-generated benchmark coverage plus bounded external consensus calibration. The verified leaderboard is stricter and only counts sourced benchmark rows. Each score includes a confidence indicator (1-4 dots) based on how much sourced coverage supports it. Display-only benchmarks — including MMLU, OpenBookQA, HumanEval, FLTEval, BBH, LisanBench, and older AIME/HMMT variants — remain visible for context but do not affect ranking.

Unless noted otherwise, ranking surfaces on this page use BenchLM's provisional leaderboard lane rather than the stricter sourced-only verified leaderboard.

Bottom line: Claude Mythos Preview leads overall, but GPT-5.4 and Claude Opus 4.6 are within striking distance — and significantly cheaper.

According to BenchLM.ai, Claude Mythos Preview leads this ranking with a score of 99, followed by Claude Opus 4.7 (97) and GPT-5.4 (93). There is meaningful separation between the top models, suggesting genuine performance differences.

The best open-weight option is GLM-5.1 (ranked #10 with a score of 84). While proprietary models lead, open-weight options are within striking distance for teams willing to trade a few points of performance for full model control.

This ranking is based on provisional overall weighted scores across BenchLM.ai's scoring formula tracked by BenchLM.ai. For detailed model profiles, click any model name below. To compare two specific models head-to-head, use the "vs #" links.

What changed

Claude Mythos Preview entered at #1 with the highest overall score on BenchLM.

GPT-5.4 holds a strong #2 across all categories.

Claude Opus 4.6 remains #3, the most consistent model across all 8 benchmark categories.

How to choose

Full Rankings (110 models)

Claude Mythos Preview
Anthropic·Proprietary·1M

99

prov. overall

Claude Opus 4.7
Anthropic·Proprietary·1M

97

prov. overall

GPT-5.4
OpenAI·Proprietary·1.05M

93

prov. overall

4
Gemini 3.1 Pro
Google·Proprietary·1M

93

prov. overall

5
GPT-5.4 Pro
OpenAI·Proprietary·1.05M

92

prov. overall

6
Claude Opus 4.6
Anthropic·Proprietary·1M

91

prov. overall

7
GPT-5.3 Codex
OpenAI·Proprietary·400K

89

prov. overall

8
Gemini 3 Pro Deep Think
Google·Proprietary·2M

86

prov. overall

9
Claude Sonnet 4.6
Anthropic·Proprietary·200K

85

prov. overall

10
GLM-5.1
Z.AI·Open Weight·203K

84

prov. overall

11
GLM-5 (Reasoning)
Z.AI·Open Weight·200K

84

prov. overall

12
GPT-5.2
OpenAI·Proprietary·400K

83

prov. overall

13
Kimi 2.6
Moonshot AI·Open Weight·256K

83

prov. overall

14
Gemini 3 Pro
Google·Proprietary·2M

83

prov. overall

15
Claude Opus 4.5
Anthropic·Proprietary·200K

80

prov. overall

16
GPT-5.1
OpenAI·Proprietary·200K

80

prov. overall

17
GPT-5.2-Codex
OpenAI·Proprietary·400K

80

prov. overall

18
Grok 4.1
xAI·Proprietary·1M

80

prov. overall

19
Qwen3.5 397B (Reasoning)
Alibaba·Open Weight·128K

80

prov. overall

20
GPT-5 (high)
OpenAI·Proprietary·128K

79

prov. overall

21
Kimi K2.5 (Reasoning)
Moonshot AI·Proprietary·128K

79

prov. overall

22
GPT-5.1-Codex-Max
OpenAI·Proprietary·400K

78

prov. overall

23
GLM-5
Z.AI·Open Weight·200K

77

prov. overall

24
Qwen3.6 Plus
Alibaba·Proprietary·1M

77

prov. overall

25
Grok 4.20
xAI·Proprietary·2M

77

prov. overall

26
GPT-5.4 mini
OpenAI·Proprietary·400K

73

prov. overall

27
GPT-5 (medium)
OpenAI·Proprietary·128K

73

prov. overall

28
Grok 4.1 Fast
xAI·Proprietary·1M

72

prov. overall

29
GLM-4.7
Z.AI·Open Weight·200K

71

prov. overall

30
Qwen3.6-35B-A3B
Alibaba·Open Weight·262K

70

prov. overall

31
Kimi K2.5
Moonshot AI·Open Weight·256K

68

prov. overall

32
Qwen3.5-122B-A10B
Alibaba·Open Weight·262K

68

prov. overall

33
Claude Sonnet 4.5
Anthropic·Proprietary·200K

68

prov. overall

34
o1-preview
OpenAI·Proprietary·200K

68

prov. overall

35
Gemini 2.5 Pro
Google·Proprietary·1M

67

prov. overall

36
Gemini 3 Flash
Google·Proprietary·1M

67

prov. overall

37
Grok 4
xAI·Proprietary·128K

67

prov. overall

38
Qwen3.5 397B
Alibaba·Open Weight·128K

66

prov. overall

39
Gemma 4 31B
Google·Open Weight·256K

66

prov. overall

40
Qwen3.5-27B
Alibaba·Open Weight·262K

65

prov. overall

41
DeepSeek V3.2 (Thinking)
DeepSeek·Open Weight·128K

65

prov. overall

42
MiniMax M2.7
MiniMax·Open Weight·200K

64

prov. overall

43
MiMo-V2-Flash
Xiaomi·Open Weight·256K

62

prov. overall

44
DeepSeek V3.2
DeepSeek·Open Weight·128K

60

prov. overall

45
GPT-4.1
OpenAI·Proprietary·1M

60

prov. overall

46
Qwen3.5-35B-A3B
Alibaba·Open Weight·262K

59

prov. overall

47
Claude Haiku 4.5
Anthropic·Proprietary·200K

59

prov. overall

48
o1
OpenAI·Proprietary·200K

59

prov. overall

49
o3
OpenAI·Proprietary·200K

59

prov. overall

50
o3-pro
OpenAI·Proprietary·200K

59

prov. overall

51
Gemma 4 26B A4B
Google·Open Weight·256K

58

prov. overall

52
o3-mini
OpenAI·Proprietary·200K

58

prov. overall

53
Claude 4.1 Opus
Anthropic·Proprietary·200K

53

prov. overall

54
DeepSeek Coder 2.0
DeepSeek·Open Weight·128K

53

prov. overall

55
DeepSeek LLM 2.0
DeepSeek·Open Weight·128K

53

prov. overall

56
Qwen2.5-1M
Alibaba·Open Weight·1M

53

prov. overall

57
Claude 4 Sonnet
Anthropic·Proprietary·200K

52

prov. overall

58
DeepSeekMath V2
DeepSeek·Open Weight·128K

52

prov. overall

59
Mistral Large 3
Mistral·Proprietary·128K

52

prov. overall

60
Qwen2.5-72B
Alibaba·Open Weight·128K

52

prov. overall

61
Gemini 3.1 Flash-Lite
Google·Proprietary·1M

51

prov. overall

62
Nemotron 3 Ultra 500B
NVIDIA·Open Weight·10M

48

prov. overall

63
Qwen3 235B 2507 (Reasoning)
Alibaba·Open Weight·128K

48

prov. overall

64
GPT-4.1 mini
OpenAI·Proprietary·1M

47

prov. overall

65
Nemotron 3 Super 100B
NVIDIA·Open Weight·1M

46

prov. overall

66
o4-mini (high)
OpenAI·Proprietary·200K

46

prov. overall

67
Claude 4.1 Opus Thinking
Anthropic·Proprietary·200K

45

prov. overall

68
GPT-4o mini
OpenAI·Proprietary·128K

45

prov. overall

69
Kimi K2
Moonshot AI·Proprietary·128K

43

prov. overall

70
Llama 3.1 405B
Meta·Open Weight·128K

43

prov. overall

71
Claude 3.5 Sonnet
Anthropic·Proprietary·200K

42

prov. overall

72
Grok Code Fast 1
xAI·Proprietary·256K

42

prov. overall

73
GPT-4o
OpenAI·Proprietary·128K

41

prov. overall

74
Sarvam 105B
Sarvam·Open Weight·128K

41

prov. overall

75
Gemini 2.5 Flash
Google·Proprietary·1M

40

prov. overall

76
Mistral Large 2
Mistral·Proprietary·128K

40

prov. overall

77
GPT-OSS 120B
OpenAI·Open Weight·128K

38

prov. overall

78
DeepSeek V3
DeepSeek·Open Weight·128K

37

prov. overall

79
Gemini 1.5 Pro
Google·Proprietary·2M

37

prov. overall

80
Claude 3 Opus
Anthropic·Proprietary·200K

36

prov. overall

81
DeepSeek-R1
DeepSeek·Open Weight·128K

35

prov. overall

82
Qwen3 235B 2507
Alibaba·Open Weight·128K

35

prov. overall

83
Grok 3 [Beta]
xAI·Proprietary·128K

34

prov. overall

84
DBRX Instruct
Databricks·Open Weight·32K

33

prov. overall

85
DeepSeek V3.1 (Reasoning)
DeepSeek·Open Weight·128K

32

prov. overall

86
o1-pro
OpenAI·Proprietary·200K

30

prov. overall

87
GLM-4.5
Z.AI·Proprietary·128K

29

prov. overall

88
Phi-4
Microsoft·Open Weight·16K

29

prov. overall

89
DeepSeek V3.1
DeepSeek·Open Weight·128K

28

prov. overall

90
GPT-4.1 nano
OpenAI·Proprietary·1M

28

prov. overall

91
Llama 3 70B
Meta·Open Weight·128K

28

prov. overall

92
GPT-4 Turbo
OpenAI·Proprietary·128K

27

prov. overall

93
Nemotron 3 Nano 30B
NVIDIA·Open Weight·32K

27

prov. overall

94
Gemini 1.0 Pro
Google·Proprietary·32K

25

prov. overall

95
Mistral 8x7B
Mistral·Open Weight·32K

25

prov. overall

96
Z-1
Z·Proprietary·128K

25

prov. overall

97
Claude 3 Haiku
Anthropic·Proprietary·200K

24

prov. overall

98
Llama 4 Scout
Meta·Open Weight·10M

24

prov. overall

99
Mixtral 8x22B Instruct v0.1
Mistral·Open Weight·64K

24

prov. overall

100
Moonshot v1
Moonshot AI·Proprietary·128K

24

prov. overall

101
Nemotron-4 15B
NVIDIA·Open Weight·32K

24

prov. overall

102
Nemotron Ultra 253B
NVIDIA·Open Weight·32K

23

prov. overall

103
GLM-4.5-Air
Z.AI·Proprietary·128K

21

prov. overall

104
GPT-OSS 20B
OpenAI·Open Weight·128K

19

prov. overall

105
Gemma 3 27B
Google·Open Weight·32K

18

prov. overall

106
Llama 4 Maverick
Meta·Open Weight·1M

18

prov. overall

107
Llama 4 Behemoth
Meta·Open Weight·32K

12

prov. overall

108
Nova Pro
Amazon·Proprietary·128K

11

prov. overall

109
Mistral 7B v0.3
Mistral·Open Weight·32K

5

prov. overall

110
Mistral 8x7B v0.2
Mistral·Open Weight·32K

2

prov. overall

These rankings update weekly

Get notified when models move. One email a week with what changed and why.

Free. No spam. Unsubscribe anytime.

Key Takeaways

The top model is Claude Mythos Preview by Anthropic with a provisional score of 99.

The best open-weight model is GLM-5.1 at position #10.

110 models are included in this ranking.

Score in Context

What these scores mean

The overall score is a weighted average across 8 benchmark categories. Agentic (22%), coding (20%), and reasoning (17%) carry the most weight. A 5-point gap in overall score is meaningful — it reflects consistent performance differences across multiple domains.

Known limitations

The overall score compresses 8 categories into one number. Two models with the same overall score can have very different strengths — one might lead coding while the other leads reasoning. Always check category scores for your specific use case.

Last updated: April 20, 2026

The AI models change fast. We track them for you.

For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.