Tokenstead - Find the right AI model for your hardware

RX 9070 XT 16GB

53 models / sorted by best_quality

Quality Fastest Least mem Newest

Model Fit Speed Memory Scores / Run

Runs comfortably

Qwen3.8-27B UD-IQ1_S Unsloth

6.2GB 14k cap

Runs comfortably

42 tok/s

ⓘ estimated

6 / 16GB

88 / 89 / 85

FIT Runs comfortably

SPEED 42 tok/s

ⓘ estimated

MEMORY 6/16GB

Qwen3.8-27B UD-Q2_K_XL Unsloth

9.8GB 9k cap

Runs comfortably

29 tok/s

ⓘ estimated

10 / 16GB

88 / 89 / 85

FIT Runs comfortably

SPEED 29 tok/s

ⓘ estimated

MEMORY 10/16GB

Ornith-1.5-9B Q8_0

9.8GB 12k cap

Runs comfortably

30 tok/s

ⓘ estimated

10 / 16GB

78 / 76 / 74

FIT Runs comfortably

SPEED 30 tok/s

ⓘ estimated

MEMORY 10/16GB

Ornith-1.5-9B Q6_K

7.6GB 16k cap

Runs comfortably

37 tok/s

ⓘ estimated

8 / 16GB

78 / 76 / 74

FIT Runs comfortably

SPEED 37 tok/s

ⓘ estimated

MEMORY 8/16GB

Ornith-1.5-9B Q5_K_M

6.6GB 18k cap

Runs comfortably

41 tok/s

ⓘ estimated

7 / 16GB

78 / 76 / 74

FIT Runs comfortably

SPEED 41 tok/s

ⓘ estimated

MEMORY 7/16GB

Ornith-1.5-9B Q4_K_M

5.8GB 20k cap

Runs comfortably

45 tok/s

ⓘ estimated

6 / 16GB

78 / 76 / 74

FIT Runs comfortably

SPEED 45 tok/s

ⓘ estimated

MEMORY 6/16GB

Phi-4 14B Q4_K_M

8.5GB 16k ctx

Runs comfortably

35 tok/s

ⓘ estimated

9 / 16GB

75 / 78 / 70

FIT Runs comfortably

SPEED 35 tok/s

ⓘ estimated

MEMORY 9/16GB

Qwen3 14B Q4_K_M

9.0GB 14k cap

Runs comfortably

32 tok/s

ⓘ estimated

9 / 16GB

72 / 75 / 70

FIT Runs comfortably

SPEED 32 tok/s

ⓘ estimated

MEMORY 9/16GB

Gemma 4 12B Q4_K_M

7.2GB 19k cap

Runs comfortably

39 tok/s

ⓘ estimated

7 / 16GB

70 / 74 / 72

FIT Runs comfortably

SPEED 39 tok/s

ⓘ estimated

MEMORY 7/16GB

Gemma 4 12B Q8_0

13.0GB 6k cap

Runs comfortably

24 tok/s

ⓘ estimated

13 / 16GB

70 / 74 / 72

FIT Runs comfortably

SPEED 24 tok/s

ⓘ estimated

MEMORY 13/16GB

Qwen3 8B Q8_0

8.5GB 15k cap

Runs comfortably

33 tok/s

ⓘ estimated

9 / 16GB

67 / 70 / 65

FIT Runs comfortably

SPEED 33 tok/s

ⓘ estimated

MEMORY 9/16GB

Qwen3 8B Q4_K_M

5.2GB 21k cap

Runs comfortably

49 tok/s

ⓘ estimated

5 / 16GB

67 / 70 / 65

FIT Runs comfortably

SPEED 49 tok/s

ⓘ estimated

MEMORY 5/16GB

Gemma 3 12B Q4_K_M

7.5GB 20k cap

Runs comfortably

38 tok/s

ⓘ estimated

8 / 16GB

62 / 65 / 68

FIT Runs comfortably

SPEED 38 tok/s

ⓘ estimated

MEMORY 8/16GB

Gemma 3 12B Q8_0

13.5GB 5k cap

Runs comfortably

23 tok/s

ⓘ estimated

14 / 16GB

62 / 65 / 68

FIT Runs comfortably

SPEED 23 tok/s

ⓘ estimated

MEMORY 14/16GB

Mistral Nemo 12B Q8_0

13.2GB 6k cap

Runs comfortably

24 tok/s

ⓘ estimated

13 / 16GB

58 / 60 / 62

FIT Runs comfortably

SPEED 24 tok/s

ⓘ estimated

MEMORY 13/16GB

Mistral Nemo 12B Q4_K_M

7.3GB 21k cap

Runs comfortably

39 tok/s

ⓘ estimated

7 / 16GB

58 / 60 / 62

FIT Runs comfortably

SPEED 39 tok/s

ⓘ estimated

MEMORY 7/16GB

Qwen3 4B Q4_K_M

2.6GB 32k ctx

Runs comfortably

94 tok/s

ⓘ estimated

3 / 16GB

58 / 60 / 56

FIT Runs comfortably

SPEED 94 tok/s

ⓘ estimated

MEMORY 3/16GB

Phi-4 Mini Q8_0

4.1GB 59k cap

Runs comfortably

72 tok/s

ⓘ estimated

4 / 16GB

55 / 58 / 52

FIT Runs comfortably

SPEED 72 tok/s

ⓘ estimated

MEMORY 4/16GB

Phi-4 Mini Q4_K_M

2.5GB 67k cap

Runs comfortably

106 tok/s

ⓘ estimated

3 / 16GB

55 / 58 / 52

FIT Runs comfortably

SPEED 106 tok/s

ⓘ estimated

MEMORY 3/16GB

Llama 3.1 8B Instruct Q8_0

8.0GB 16k cap

Runs comfortably

35 tok/s

ⓘ estimated

8 / 16GB

52 / 55 / 62

FIT Runs comfortably

SPEED 35 tok/s

ⓘ estimated

MEMORY 8/16GB

Llama 3.1 8B Instruct Q4_K_M

4.5GB 23k cap

Runs comfortably

54 tok/s

ⓘ estimated

5 / 16GB

52 / 55 / 62

FIT Runs comfortably

SPEED 54 tok/s

ⓘ estimated

MEMORY 5/16GB

Gemma 3 4B Q8_0

5.0GB 36k cap

Runs comfortably

57 tok/s

ⓘ estimated

5 / 16GB

42 / 45 / 55

FIT Runs comfortably

SPEED 57 tok/s

ⓘ estimated

MEMORY 5/16GB

Gemma 3 4B Q4_K_M

3.0GB 43k cap

Runs comfortably

83 tok/s

ⓘ estimated

3 / 16GB

42 / 45 / 55

FIT Runs comfortably

SPEED 83 tok/s

ⓘ estimated

MEMORY 3/16GB

Llama 3.2 3B Instruct Q8_0

3.5GB 52k cap

Runs comfortably

79 tok/s

ⓘ estimated

4 / 16GB

35 / 38 / 50

FIT Runs comfortably

SPEED 79 tok/s

ⓘ estimated

MEMORY 4/16GB

Llama 3.2 3B Instruct Q4_K_M

2.0GB 58k cap

Runs comfortably

118 tok/s

ⓘ estimated

2 / 16GB

35 / 38 / 50

FIT Runs comfortably

SPEED 118 tok/s

ⓘ estimated

MEMORY 2/16GB

Mellum2.1 12B Thinking Q4_K_M

7.0GB 109k cap

Runs comfortably

212 tok/s

ⓘ estimated

7 / 16GB

FIT Runs comfortably

SPEED 212 tok/s

ⓘ estimated

MEMORY 7/16GB

Runs tight

Gemma 4 26B A4B Q4_K_M

13.0GB ctx headroom ~0

Runs tight

73 tok/s

ⓘ estimated

13 / 16GB

85 / 88 / 84

FIT Runs tight

SPEED 73 tok/s

ⓘ estimated

MEMORY 13/16GB

Qwen3 30B A3B Q4_K_M

15.5GB ctx headroom ~0

Runs tight

93 tok/s

ⓘ estimated

16 / 16GB

75 / 78 / 72

FIT Runs tight

SPEED 93 tok/s

ⓘ estimated

MEMORY 16/16GB

Phi-4 14B Q8_0

15.0GB 2k cap

Runs tight

21 tok/s

ⓘ estimated

15 / 16GB

75 / 78 / 70

FIT Runs tight

SPEED 21 tok/s

ⓘ estimated

MEMORY 15/16GB

Qwen3 14B Q8_0

15.0GB 2k cap

Runs tight

21 tok/s

ⓘ estimated

15 / 16GB

72 / 75 / 70

FIT Runs tight

SPEED 21 tok/s

ⓘ estimated

MEMORY 15/16GB

Llama 3.1 8B Instruct FP16

16.0GB ctx headroom ~0

Runs tight

20 tok/s

ⓘ estimated

16 / 16GB

52 / 55 / 62

FIT Runs tight

SPEED 20 tok/s

ⓘ estimated

MEMORY 16/16GB

LFM2.5-24B-A2B Q4_K_M

13.7GB ctx headroom ~0

Runs tight

241 tok/s

ⓘ estimated

14 / 16GB

FIT Runs tight

SPEED 241 tok/s

ⓘ estimated

MEMORY 14/16GB

Runs with CPU offload

Qwen3.8-Flash-Next EXL3_2BPW

62.7GB ctx spills to ram

Runs with CPU offload

0.5 tok/s

ⓘ estimated

63 / 16GB

92 / 92 / 89

FIT Runs with CPU offload

SPEED 0.5 tok/s

ⓘ estimated

MEMORY 63/16GB

Qwen3.8-Flash-Next EXL3_3BPW

85.0GB ctx spills to ram

Runs with CPU offload

0.5 tok/s

ⓘ estimated

85 / 16GB

92 / 92 / 89

FIT Runs with CPU offload

SPEED 0.5 tok/s

ⓘ estimated

MEMORY 85/16GB

Ornith-1.5-35B-A3B Q4_K_M

21.7GB ctx spills to ram

Runs with CPU offload

1 tok/s

ⓘ estimated

22 / 16GB

91 / 89 / 86

FIT Runs with CPU offload

SPEED 1 tok/s

ⓘ estimated

MEMORY 22/16GB

Ornith-1.5-35B-A3B Q5_K_M

25.3GB ctx spills to ram

Runs with CPU offload

0.83 tok/s

ⓘ estimated

25 / 16GB

91 / 89 / 86

FIT Runs with CPU offload

SPEED 0.83 tok/s

ⓘ estimated

MEMORY 25/16GB

Qwen3.6 35B A3B Q4_K_M

20.0GB ctx spills to ram

Runs with CPU offload

2 tok/s

ⓘ estimated

20 / 16GB

90 / 91 / 87

FIT Runs with CPU offload

SPEED 2 tok/s

ⓘ estimated

MEMORY 20/16GB

Pokee-Isaac 28B Q4_K_M

16.0GB ctx spills to ram

Runs with CPU offload

9 tok/s

ⓘ estimated

16 / 16GB

90 / 88 / 88

FIT Runs with CPU offload

SPEED 9 tok/s

ⓘ estimated

MEMORY 16/16GB

Qwen3.8-27B UD-Q4_K_XL Unsloth

17.6GB ctx spills to ram

Runs with CPU offload

7 tok/s

ⓘ estimated

18 / 16GB

88 / 89 / 85

FIT Runs with CPU offload

SPEED 7 tok/s

ⓘ estimated

MEMORY 18/16GB

Gemma 4 31B QAT

18.0GB ctx spills to ram

Runs with CPU offload

8 tok/s

ⓘ estimated

18 / 16GB

86 / 89 / 85

FIT Runs with CPU offload

SPEED 8 tok/s

ⓘ estimated

MEMORY 18/16GB

Gemma 4 31B Q4_K_M

18.3GB ctx spills to ram

Runs with CPU offload

7 tok/s

ⓘ estimated

18 / 16GB

86 / 89 / 85

FIT Runs with CPU offload

SPEED 7 tok/s

ⓘ estimated

MEMORY 18/16GB

Qwen3.6 27B Q4_K_M

16.5GB ctx spills to ram

Runs with CPU offload

13 tok/s

ⓘ estimated

17 / 16GB

85 / 87 / 82

FIT Runs with CPU offload

SPEED 13 tok/s

ⓘ estimated

MEMORY 17/16GB

Qwen3 32B Q4_K_M

20.0GB ctx spills to ram

Runs with CPU offload

7 tok/s

ⓘ estimated

20 / 16GB

80 / 82 / 78

FIT Runs with CPU offload

SPEED 7 tok/s

ⓘ estimated

MEMORY 20/16GB

Qwen3 32B Q5_K_M

24.5GB ctx spills to ram

Runs with CPU offload

5 tok/s

ⓘ estimated

25 / 16GB

80 / 82 / 78

FIT Runs with CPU offload

SPEED 5 tok/s

ⓘ estimated

MEMORY 25/16GB

Ornith-1.5-9B BF16

18.4GB ctx spills to ram

Runs with CPU offload

8 tok/s

ⓘ estimated

18 / 16GB

78 / 76 / 74

FIT Runs with CPU offload

SPEED 8 tok/s

ⓘ estimated

MEMORY 18/16GB

Xing4.0-29B-A4B Q4_K_M

20.1GB ctx spills to ram

Runs with CPU offload

3 tok/s

ⓘ estimated

20 / 16GB

75 / 73 / 72

FIT Runs with CPU offload

SPEED 3 tok/s

ⓘ estimated

MEMORY 20/16GB

Llama 3.3 70B Instruct Q2_K

21.5GB ctx spills to ram

Runs with CPU offload

6 tok/s

ⓘ estimated

22 / 16GB

72 / 77 / 80

FIT Runs with CPU offload

SPEED 6 tok/s

ⓘ estimated

MEMORY 22/16GB

Gemma 3 27B Q4_K_M

17.0GB ctx spills to ram

Runs with CPU offload

11 tok/s

ⓘ estimated

17 / 16GB

68 / 71 / 72

FIT Runs with CPU offload

SPEED 11 tok/s

ⓘ estimated

MEMORY 17/16GB

Mixtral 8x7B Instruct Q4_K_M

26.5GB ctx spills to ram

Runs with CPU offload

2 tok/s

ⓘ estimated

27 / 16GB

62 / 65 / 68

FIT Runs with CPU offload

SPEED 2 tok/s

ⓘ estimated

MEMORY 27/16GB

Mixtral 8x7B Instruct Q3_K_M

19.0GB ctx spills to ram

Runs with CPU offload

4 tok/s

ⓘ estimated

19 / 16GB

62 / 65 / 68

FIT Runs with CPU offload

SPEED 4 tok/s

ⓘ estimated

MEMORY 19/16GB

Mellum2.1 12B Thinking FP16

24.3GB ctx spills to ram

Runs with CPU offload

0.99 tok/s

ⓘ estimated

24 / 16GB

FIT Runs with CPU offload

SPEED 0.99 tok/s

ⓘ estimated

MEMORY 24/16GB

Muse Glimmer 30B Q4_K_M

18.0GB ctx spills to ram

Runs with CPU offload

4 tok/s

ⓘ estimated

18 / 16GB

FIT Runs with CPU offload

SPEED 4 tok/s

ⓘ estimated

MEMORY 18/16GB

LFM2.5-24B-A2B FP8

24.0GB ctx spills to ram

Runs with CPU offload

0.5 tok/s

ⓘ estimated

24 / 16GB

FIT Runs with CPU offload

SPEED 0.5 tok/s

ⓘ estimated

MEMORY 24/16GB

Coding alternatives

Hosted coding assistants for comparison - per seat

Cursor Hobby

AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...

$0.0

/MONTH

GitHub Copilot Individual

AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...

$10.0

/MONTH

GitHub Copilot Business

AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...

$19.0

/MONTH

Cursor Pro

AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...

$20.0

/MONTH

GitHub Copilot Enterprise

AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...

$39.0

/MONTH

Cursor Business

AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...

$40.0

/MONTH

Per-token monthly figures are estimates at typical usage. Prices verified periodically - always confirm current pricing on the provider's website.

On a monthly budget? Compare owning hardware vs renting cloud →

Tokenstead community

Not sure this is the right fit?

Post your rig and the shortlist in the community and get a sanity check from people running the same memory and stack.

Join the community Share your rig