Model comparison
Mercury 2 vs GPT-5
Capability radar, value bars, and side-by-side posture.
Mercury 2
Not scored
value score
GPT-5
94
value score
Mercury 2 ctx
—
GPT-5 ctx
400K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- GPT-594
Attribute tape
| Field | Mercury 2 | GPT-5 |
|---|---|---|
| Developer | Inception | OpenAI |
| Context | — | 400K |
| Modalities | Text, Code, Image, Audio | |
| Openness | — | Proprietary Model |
| Speed | — | Fast |
| Price | — | Mid-High |
| Value score | — | 94 |
| Summary | Mercury 2 is Inception Labs' diffusion-based reasoning language model that uses parallel token refinement instead of left-to-right autoregressive generation to achieve very low latency inference. It targets production AI workloads such as coding assistants, agents, voice pipelines, and retrieval-augmented applications. | OpenAI frontier multimodal model. |