Mistral Large 3 vs GLM-5.3

Mistral AI's Mistral Large 3 and Zhipu AI's GLM-5.3 compared on specs, pricing, and capabilities — as configured in AI Crucible ensembles.

Mistral Large 3 input tokens are 2.8x cheaper than GLM-5.3 ($0.60 vs $1.68 per 1M tokens).
Mistral Large 3 output tokens are 2.9x cheaper than GLM-5.3 ($1.80 vs $5.28 per 1M tokens).
GLM-5.3 offers the larger context window: 1M tokens vs 256K tokens for Mistral Large 3.
GLM-5.3 is a reasoning model that spends extra tokens thinking before answering; Mistral Large 3 responds directly.
Mistral Large 3 accepts image inputs; GLM-5.3 is text-only in AI Crucible.

Specification	Mistral Large 3	GLM-5.3
Provider	Mistral AI	Zhipu AI
API model ID	mistral-large-latest	zai-org/GLM-5.3
Description	Mistral AI's flagship model with 41B active parameters, excelling at complex reasoning and multimodal tasks.	Z.AI's frontier open-weight model for coding and agentic work, with always-on reasoning and 1M context.
Context window	256K tokens	1M tokens
Max output tokens	8,192	131,072
Input cost ($/1M tokens)	$0.60	$1.68
Output cost ($/1M tokens)	$1.80	$5.28
Cache read cost ($/1M tokens)	-	$0.31
Latency	medium	medium
Reasoning model	No	Yes
Vision (image input)	Yes	No
Tool use	Yes	Yes