Best Non-Reasoning LLMs in 2026

Top standard AI models (no chain-of-thought reasoning) ranked by benchmark performance. Faster and cheaper than reasoning models.

Claude Opus 4.6
AnthropicProprietary1M

90

overall

Gemini 3.1 Pro
GoogleProprietary1M

89

overall

Grok 4.1
xAIProprietary128K

89

overall

4
Claude Sonnet 4.6
AnthropicProprietary1M

86

overall

5
Claude Opus 4.5
AnthropicProprietary200K

85

overall

6
Gemini 3 Pro
GoogleProprietary2M

84

overall

7
Claude Sonnet 4.5
AnthropicProprietary1M

83

overall

8
Grok 4.1 Fast
xAIProprietary2M

83

overall

9
Grok 4
xAIProprietary128K

74

overall

10
GLM-5
Zhipu AIOpen Weight200K

74

overall

11
DeepSeek Coder 2.0
DeepSeekOpen Weight128K

73

overall

12
Qwen2.5-1M
AlibabaOpen Weight1M

73

overall

13
Qwen2.5-72B
AlibabaOpen Weight128K

72

overall

14
DeepSeek V3.2
DeepSeekOpen Weight128K

72

overall

15
Gemini 2.5 Pro
GoogleProprietary2M

72

overall

16
o4-mini (high)
OpenAIProprietary200K

71

overall

17
Qwen3.5 397B
AlibabaOpen Weight128K

71

overall

18
DeepSeek LLM 2.0
DeepSeekOpen Weight128K

70

overall

19
Claude 4.1 Opus
AnthropicProprietary200K

69

overall

20
Kimi K2.5
Moonshot AIOpen Weight128K

68

overall

21
Mistral Large 3
MistralOpen Weight128K

68

overall

22
Claude 4 Sonnet
AnthropicProprietary200K

67

overall

23
MiniMax M2.5
MiniMaxProprietary128K

66

overall

24
Llama 3.1 405B
MetaOpen Weight128K

65

overall

25
Gemini 3 Flash
GoogleProprietary1M

65

overall

26
Mistral Large 2
MistralProprietary128K

64

overall

27
Claude Haiku 4.5
AnthropicProprietary200K

64

overall

28
Nemotron 3 Super 100B
NVIDIAOpen Weight32K

62

overall

29
Claude 3.5 Sonnet
AnthropicProprietary200K

62

overall

30
GPT-4o
OpenAIProprietary128K

60

overall

31
Grok Code Fast 1
xAIProprietary256K

58

overall

32
Mistral 8x7B
MistralOpen Weight32K

57

overall

33
Gemini 1.5 Pro
GoogleProprietary2M

55

overall

34
Gemini 3.1 Flash-Lite
GoogleProprietary1M

55

overall

35
Claude 3 Opus
AnthropicProprietary200K

53

overall

36
GPT-4 Turbo
OpenAIProprietary128K

53

overall

37
Gemini 1.0 Pro
GoogleProprietary32K

53

overall

38
Nemotron-4 15B
NVIDIAOpen Weight32K

52

overall

39
Nemotron 3 Nano 30B
NVIDIAOpen Weight32K

52

overall

40
Llama 3 70B
MetaOpen Weight128K

51

overall

41
Claude 3 Haiku
AnthropicProprietary200K

51

overall

42
Moonshot v1
Moonshot AIProprietary128K

51

overall

43
Z-1
ZProprietary128K

51

overall

44
GPT-OSS 120B
OpenAIOpen Weight128K

50

overall

45
Gemini 2.5 Flash
GoogleProprietary1M

49

overall

46
Llama 4 Behemoth
MetaOpen Weight32K

44

overall

47
Llama 4 Scout
MetaOpen Weight32K

42

overall

48
Llama 4 Maverick
MetaOpen Weight32K

42

overall

49
Gemma 3 27B
GoogleOpen Weight32K

42

overall

50
Qwen2.5-VL-32B
AlibabaOpen Weight32K

41

overall

51
Grok 3 [Beta]
xAIProprietary128K

40

overall

52
Nova Pro
Nova AIProprietary128K

40

overall

53
Qwen3 235B 2507
AlibabaOpen Weight128K

38

overall

54
GLM-4.5
TsinghuaProprietary128K

38

overall

55
MiniMax M1 80k
MiniMaxProprietary80K

37

overall

56
GLM-4.5-Air
TsinghuaProprietary128K

37

overall

57
DeepSeek V3.1
DeepSeekOpen Weight128K

35

overall

58
Kimi K2
Moonshot AIProprietary128K

35

overall

59
GPT-OSS 20B
OpenAIOpen Weight128K

34

overall

60
Mistral 7B v0.3
MistralOpen Weight32K

34

overall

61
Mistral 8x7B v0.2
MistralOpen Weight32K

33

overall

Key Takeaways

  • The top model is Claude Opus 4.6 by Anthropic with a score of 90.
  • The best open-weight model in this ranking is GLM-5 at position #10.
  • 61 models are included in this ranking.