Mercury 2.5

Mercury 2.5 is a reasoning model from Inception. 14 benchmarks count toward its score, in 6 categories.

availableShows if the model has enough results for an index.
IndexOverall score out of 100.48.5 ±5.8
CoverageShare of the index weight with results.85%
SpeedOutput tokens per second.196/s
Input / 1MUS dollars per 1M input tokens.$0.04
Output / 1MUS dollars per 1M output tokens.$0.15
ContextMaximum tokens in one request.260K
EloLMArena rating and rank.N/A

The index is a score out of 100. The ± range shows how much it can change.

CapabilitiesScore per category, out of 100.

Out of 100
AgenticMulti-step tasks with tools.
51.3
CodingCode writing and repair.
44.9
ReasoningLogic problems and puzzles.
55.3
MultimodalTasks with images and text.
N/A
KnowledgeFacts and expert knowledge.
46.1
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
55.6
MathMath problems.
50.6

Results

14 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result as a score out of 100.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
τ³-Bench Tool-Agent-User EvaluationAgentic96.0%76.7Sierra Research
GPQA DiamondKnowledge79.0%51.2David Rein et al.
Instruction Following BenchmarkInstruction77.0%55.6Benchmark authors
Artificial Analysis Long Context ReasoningReasoning68.0%55.3Artificial Analysis
Scientific Code BenchmarkCoding38.0%47.8Benchmark authors
Terminal-Bench 2.1Agentic34.1%44.221 Sept 2026Vals AI
DeepSearchQAAgentic34.0%24.7Meta AI
Artificial Analysis Omniscience AccuracyKnowledge22.0%41.1Artificial Analysis
SkillsBenchCoding18.1%38.3OpenHands11 Sept 2026Vals AI
Code MigrationCoding4.5%47.821 Sept 2026Vals AI
Vibe Code Bench v1.1Coding3.7%43.7OpenHands21 Sept 2026Vals AI
ProofBench v1.1Math3.0%50.621 Sept 2026Vals AI
IOICoding2.4%46.721 Sept 2026Vals AI
Terminal-Bench 4.0Agentic0.0%59.521 Sept 2026Vals AI

14 benchmarks count, from 14 of 14 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AI

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from Inception

Mercury 2UnrankedMercury 2.5 PreviewUnrankedMercury Edit 2Unranked