BenchLeader
Menu
LeaderboardsModelsProvidersBenchmarksCompareReleasesBlogMethodology
InceptionReasoning model

Mercury 2

Best configuration ranks #306 of 610 on the BenchLeader Index at 49.5 ±6.7. Last measured 8 Sept 2026. Released 24 Feb 2026.

Blended price
$0.375/M
$0.250 in · $0.750 out
Output speed
744 tok/s
First answer
4.04 s
first token 0.86 s
Context
128k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index50
  2. Reasoning52
  3. Coding36
  4. Agents & tools57
  5. Knowledge40
  6. Instruction following68
  7. Human preference52
  8. Long context48
  9. Composite44

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkdefaultSourceTrend
LMArena Hard Prompts1360#185LMArena
GPQA Diamond (AA)not in index77.0%#218Artificial Analysis
Humanity's Last Exam (AA)not in index17.1%#179Artificial Analysis

Coding

BenchmarkdefaultSourceTrend
SciCode38.7%#118SciCode
WeirdML43.2%#88WeirdML
LMArena Coding1394#177LMArena
LMArena WebDev1166#115LMArena
SciCode (AA)not in index37.7%#128Artificial Analysis

Agents & tools

BenchmarkdefaultSourceTrend
Terminal-Bench Hard26.5%#126Artificial Analysis
τ²-Bench Telecom (AA)not in index70.8%#149Artificial Analysis

Knowledge

BenchmarkdefaultSourceTrend
AA-Omniscience-50.7#318Artificial Analysis

Instruction following

BenchmarkdefaultSourceTrend
IFBench69.8%#71Artificial Analysis

Human preference

BenchmarkdefaultSourceTrend
LMArena Text1347#184LMArena

Long context

BenchmarkdefaultSourceTrend
AA-LCR43.7%#274Artificial Analysis

Composite

BenchmarkdefaultSourceTrend
AA Intelligence Index11.5#283Artificial Analysis

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Inception253 tok/s0.86 s$0.250$0.750128k–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.025 per 1M. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0003$0.00034.4 s
Summarise a 30-page report12,000 / 600$0.0035$0.00144.8 s
Code edit6,000 / 1,500$0.0026$0.00166.1 s
Agentic coding session60,000 / 4,000$0.018$0.00799.4 s
Structured extraction2,000 / 200$0.0007$0.00034.3 s

See also

Best under $2/MBest under $0.50/MBest for codingAll Inception models

Data as of 9 Sept 2026. Compare with another model.