The budget tier, where blended prices sit under a dollar per million tokens. Capability stays visible beside every price so cheap never hides weak.Last reviewed 2026-09-09.
Z.ai's efficient multimodal model for fast coding and long-horizon agent tasks.
MiniMax's multimodal foundation model for long-horizon agentic work.
Efficiency-optimized DeepSeek V4 at 284B total / 13B active for fast, high-throughput inference.
No shared demos for these models yet.
No captured outputs for these models yet.
DeepSeek V4 Flash has the lowest blended list price. Per 1M tokens: GLM 5.3 Flash at $0.15 in / $0.50 out; MiniMax M3 at $0.30 in / $1.20 out; DeepSeek V4 Flash at $0.09 in / $0.18 out.
GLM 5.3 Flash leads with 1M tokens. The other two: MiniMax M3 at 524K, DeepSeek V4 Flash at 1M.
GLM 5.3 Flash and MiniMax M3 accept images; the third do not.
This idapt comparison runs all three in aligned columns: the same prompt, plus benchmarks, specs, and prices per model. Open any model in chat from its column to keep working with it.