Method

Kimi K3.1 Pro vs GLM-5.3: unrestricted models compared

Models compared

Side by side

Green marks the better value in each row. Read November 18, 2026.

Kimi K3.1 ProGLM-5.3
Refusals
How refusals were removedAbliteratedAbliterated
Size and context
Parameterstotal2,780B753B
Active parametersper tokenAllAll
Context windowtokens1,048,5761,048,576
Max outputtokens943,718131,072
Inputs
Text
Images
Video
Price, lowest listed host
InputUSD per 1M$3.50$2.00
OutputUSD per 1M$8.00$4.00
BlendedUSD per 1M$4.63$2.50
Where it runs
HostsSilk ComputeSilk Compute, Abliteration
API formatsat the host3 of 33 of 3
Runs locally

Which to pick

Kimi K3.1 Pro

Pick Kimi K3.1 Pro for security work, images or video, and outputs longer than 131,072 tokens.

GLM-5.3

Pick GLM-5.3 for text-only software work at $2.00 in and $4.00 out.

Against the other models

Size

Total parameters, billions. Log scale. Larger is more capable

2,780 Kimi K3.1 Pro 763 DeepSeek V4.1 Flash 753 GLM-5.3 405 Hermes 3 405B 320 GLM-5.3 Flash 36 Qwen 3.6 35B 31.3 Gemma 4 31B 27.8 Qwen 3.8 27B 8 Llama 3.1 8B

Context window

Tokens the model holds at once. Higher is better

1M Kimi K3.1 Pro 1M DeepSeek V4.1 Flash 1M GLM-5.3 1M GLM-5.3 Flash 256K Qwen 3.6 35B 256K Gemma 4 31B 256K Qwen 3.8 27B 128K Hermes 3 405B 128K Llama 3.1 8B

Price

Blended USD per 1M tokens, lowest listed host. Lower is better

$0.41 DeepSeek V4.1 Flash $0.43 GLM-5.3 Flash $1.20 Hermes 3 405B $2.50 GLM-5.3 $4.20 Qwen 3.8 27B $4.63 Kimi K3.1 Pro
AbliteratedFine-tunedHatched: weights only, no API listed