Inception model product
Mercury 2 is the fastest reasoning LLM, built on diffusion-based language model (dLLM) architecture. Instead of generating text token-by-token, it refines multiple text blocks simultaneously, achieving over 1,000 tokens per second on Nvidia Blackwell GPUs — 5x faster than leading speed-optimized LLMs. Supports tool usage and JSON output with 128K context window.
Updated Aug 10, 2026. Default version: Mercury 2
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Mercury 2.
| IFBench | 0.7 | 14 | 28 | 51.9% | C | |
| SciCode | 0.4 | 15 | 18 | 17.6% | C | |
| Tau2 Airline | 0.5 | 19 | 23 | 18.2% | C | |
| LiveCodeBench | 0.7 | 24 | 73 | 68.1% | C | |
| AIME 2025 | 0.9 | 43 | 114 | 62.8% | C | |
| GPQA | 0.7 | 113 | 233 | 51.7% | C |
Preference and agent-evaluation signals from published Arena datasets.
| webdev | webdev-react | 93 | 1155.9 | 784 | N/A | |
| webdev | webdev-html | 106 | 1198.5 | 103 | N/A | |
| webdev | overall | 107 | 1166.3 | 888 | N/A | |
| webdev | webdev | 107 | 1166.3 | 888 | N/A | |
| text factuality | english | 138 | 1393.8 | 1,377 | N/A | |
| text factuality | non english | 141 | 1349.9 | 1,721 | N/A | |
| text factuality | overall | 144 | 1372.2 | 3,098 | N/A | |
| text factuality | hard prompts | 144 | 1387.1 | 1,751 | N/A | |
| text factuality | exclude ties | 145 | 1341.4 | 2,155 | N/A | |
| text | coding | 152 | 1393.0 | 764 | N/A | |
| text | hard prompts english | 160 | 1378.0 | 855 | N/A | |
| text | industry software and it services | 160 | 1386.5 | 1,081 | N/A | |
| text | industry medicine and healthcare | 161 | 1369.9 | 227 | N/A | |
| text | industry business and management and financial operations | 162 | 1351.6 | 481 | N/A | |
| text | english | 164 | 1379.2 | 1,378 | N/A | |
| text | expert | 165 | 1355.6 | 229 | N/A | |
| text | multi turn | 168 | 1355.6 | 536 | N/A | |
| text | hard prompts | 170 | 1362.6 | 1,751 | N/A | |
| text | overall | 172 | 1357.6 | 3,099 | N/A | |
| text | industry entertainment and sports and media | 172 | 1318.9 | 653 | N/A | |
| text | exclude ties | 175 | 1315.5 | 2,156 | N/A | |
| text | non english | 175 | 1332.3 | 1,721 | N/A | |
| text | instruction following | 182 | 1322.9 | 834 | N/A | |
| text style control | coding | 187 | 1396.4 | 764 | N/A | |
| text | longer query | 188 | 1327.9 | 838 | N/A | |
| text | industry life and physical and social science | 191 | 1346.3 | 469 | N/A | |
| text style control | expert | 192 | 1347.4 | 229 | N/A | |
| text style control | industry software and it services | 192 | 1384.7 | 1,081 | N/A | |
| text style control | industry entertainment and sports and media | 193 | 1313.2 | 653 | N/A | |
| text style control | multi turn | 195 | 1341.8 | 536 | N/A | |
| text | creative writing | 196 | 1295.2 | 524 | N/A | |
| text style control | non english | 196 | 1325.0 | 1,721 | N/A | |
| text | industry writing and literature and language | 197 | 1303.9 | 688 | N/A | |
| text style control | overall | 197 | 1346.6 | 3,099 | N/A | |
| text style control | english | 197 | 1364.8 | 1,378 | N/A | |
| text | russian | 198 | 1303.8 | 358 | N/A | |
| text style control | hard prompts english | 198 | 1372.1 | 855 | N/A | |
| text style control | hard prompts | 199 | 1359.8 | 1,751 | N/A | |
| text style control | industry business and management and financial operations | 202 | 1340.7 | 481 | N/A | |
| text style control | exclude ties | 203 | 1299.5 | 2,156 | N/A | |
| text style control | instruction following | 205 | 1322.9 | 834 | N/A | |
| text style control | industry medicine and healthcare | 207 | 1345.3 | 227 | N/A | |
| text style control | creative writing | 215 | 1299.3 | 524 | N/A | |
| text style control | industry writing and literature and language | 215 | 1310.1 | 688 | N/A | |
| text style control | russian | 226 | 1304.9 | 358 | N/A | |
| text style control | longer query | 234 | 1323.6 | 838 | N/A | |
| text style control | industry life and physical and social science | 246 | 1323.4 | 469 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| NanoGPT | mercury-2 | global | $0.25 | $0.75 | 128K | |
| Inception | mercury-2 | global | $0.25 | $0.75 | 128K | |
| Vercel AI Gateway | inception/mercury-2 | global | $0.25 | $0.75 | 128K | |
| OpenRouter | inception/mercury-2 | global | $0.25 | $0.75 | 128K | |
| Venice AI | mercury-2 | global | $0.3125 | $0.9375 | 128K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Mercury 2 | 50.8 | N/A | 128K | 8.2K | No | Proprietary |
A concise description based on the published model registry.
Mercury 2 is the fastest reasoning LLM, built on diffusion-based language model (dLLM) architecture. Instead of generating text token-by-token, it refines multiple text blocks simultaneously, achieving over 1,000 tokens per second on Nvidia Blackwell GPUs — 5x faster than leading speed-optimized LLMs. Supports tool usage and JSON output with 128K context window.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Mercury 2.
Mercury 2's default version was released on Feb 24, 2026.
Mercury 2's official API price is $0.25 per million input tokens and $0.75 per million output tokens via Inception. The lowest tracked third-party offer starts at $0.25 input and $0.75 output via NanoGPT.
Mercury 2 is published under Inception in the model registry.
The default version has a 128K token context window.
No. The default version is not marked as having publicly available weights.
5 published provider offerings are linked to the default version.
Nearby ranked alternatives include Qwen2.5 VL 32B, Kimi K2, GPT-4-Turbo.