Mercury 2.5

Released2026-09-08
Price per 1M$0.20 in · $0.75 out
Overall score45.4 · rank 154
Context window260,000 tokens
ProviderInception
WeightsProprietary

Scores

DimensionScoreRankBenchmarks
Overall: Balanced 45.4 154 12
Overall: Coding 45.6 148 10
Overall: Reasoning 43.0 171 3
Overall: Agentic work 48.9 121 9
Overall: Frontend 39.8 172 3
Coding 46.8 140 2
Agents 51.8 96 6
Reasoning 40.8 189 1
Math 50.3 110 1
Knowledge 35.4 153 1
Frontend 36.3 119 1
Value vs price+0.8114

Configurations

ConfigurationBalancedBenchmarks
Unspecified setting 44.6 ±4.3 3
High 47.0 ±2.8 10

Evidence

Coding 46.8 ±3.1

BenchmarkResultSourceDate
IOI 2.4% Vals AI 2026-10-29
SciCode 38.5% Epoch AI 2026-11-11

Agents 51.8 ±3.7

BenchmarkResultSourceDate
Terminal-Bench 4.0 (Vals run) 0.0% Vals AI 2026-11-09
Terminal-Bench-Science 0.0% Vals AI 2026-11-09
Terminal-Bench 2.1 (Vals run) 34.5% Vals AI 2026-10-01
Tax Agent Bench 2.5% Vals AI 2026-11-09
SkillsBench 18.1% Vals AI 2026-10-29
CyberBench v1.1 45.5% Vals AI 2026-11-05

Reasoning 40.8 ±5.3

BenchmarkResultSourceDate
CritPt 0.0% Epoch AI 2026-11-11

Math 50.3 ±2.9

BenchmarkResultSourceDate
ProofBench v1.1 3.0% Vals AI 2026-11-09
ProofBench 3.0% Epoch AI 2026-11-11

Knowledge 35.4 ±6.5

BenchmarkResultSourceDate
Legal Research Bench 4.3% Vals AI 2026-11-09

Frontend 36.3 ±0.8

BenchmarkResultSourceDate
Vibe Code Bench v1.1 3.7% Vals AI 2026-11-09

Names matched to this model

mercury-2-5

Compare Mercury 2.5 with other models