Mercury 2
Inception released 2026-03-04
$0.375 per million tokens, balanced
DeepSeek V4 Flash 0731 is both better and cheaper.
It scores +29.9 higher and costs 72% less ($0.105/M against $0.375/M) on this workload — and it does everything this model does.
Ling-3.0-flash is cheaper still (92% less) but drops 50K → 33K max output.
Every price dimension
| Input | $0.250/M |
|---|---|
| Output | $0.750/M |
| Cached input | $0.025/M |
Independent scores
| Intelligence | 21.9 |
|---|---|
| Coding | 31.1 |
| Agentic | 9.5 |
| LMArena Elo | 1357.8default |
Sources are listed separately rather than averaged. Across the 62 models both have scored they correlate at r = 0.806 — close agreement overall, but four models rank very differently between them, and a blended score would hide exactly those.
Capability
| Context window | 128K |
|---|---|
| Max output | 50K |
| Input modes | text |
| Tool use | yes |
| Reasoning | optional |
Provenance
| Price source | openrouter.ai |
|---|---|
| Fetched | 2026-08-24 |
| Quality data | verified |
Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM). Instead of generating tokens sequentially, Mercury 2 produces and refines multiple tokens in parallel, achieving...