Inference
EN
Inception

Mercury 3

Mercury 3 is a reasoning model from Inception. 8 benchmarks count toward its index, in 5 categories.

  • Closed weights
  • Reasoning
  • Input: text
  • Released 2026-10
  • Retires 8 Nov 2026
RankedEnough results for an index.
IndexAll categories, out of 100.52.3±8.2Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.295/sMedian 67/s
First tokenSeconds to the first output token.0.3 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.10Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.35Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
inception/mercury-3
Max outputMaximum output tokens per request.
131K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.01
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
8 Nov 2026

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted44.4164 / 218
CodingCode writing and repair.1 counted64.2—
ReasoningLogic problems and puzzles.1 counted57.0—
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.1 counted59.0—
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted60.8—
MathMath problems.None counted——

Results

8 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Instruction Following BenchmarkInstruction76.8%24 / 57
GPQA Diamond · BenchLMKnowledge89.2%37 / 84
Vibe Code Bench v1.1Coding60.9% ±5.459 / 123
MRCRv2Reasoning75.5%7 / 14
τ³-Bench Tool-Agent-User EvaluationAgentic30.2%25 / 32
Terminal-Bench 4.0Agentic3.0%53 / 61
DeepSearchQAAgentic71.3%21 / 24
DeepSWEAgentic46.3%52 / 59
Legal Research BenchLegal28.8% ±3.157 / 90
Tax Agent BenchFinance18.2% ±2.556 / 82
HLAB-Vals-1.1Legal5.0% ±1.627 / 35
MultiChallenge: provider-reported runsInstruction60.4%—
BrowseComp-PlusAgentic55.2%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseVals AINo licence stated

History

5 releases
Index52.7 → 52.3
29 Oct 202612 Nov 2026
Output / 1M$0.35 → $0.35
23 Oct 202612 Nov 2026

Questions

How much does Mercury 3 cost?
$0.10 per 1M input tokens, $0.35 per 1M output tokens.
How good is Mercury 3?
Its index is 52.3 out of 100. That is place 124 of 283 ranked models.
What is Mercury 3 best at?
Agentic: 44.4 out of 100, place 164 of 218.
How fast is Mercury 3?
295 output tokens per second, 0.3 s to the first token.
What is the context window of Mercury 3?
1M tokens.
Is Mercury 3 open source?
No. Its weights are not published.
When was Mercury 3 released?
October 2026.

Badge

Mercury 3 on Inference360: place 124 of 283, index 52.3
HTML
<a href="https://inference360.ai/models/mercury-3"><img src="https://inference360.ai/badges/mercury-3.svg" alt="Mercury 3 on Inference360: place 124 of 283, index 52.3" height="20"></a>
Markdown
[![Mercury 3 on Inference360: place 124 of 283, index 52.3](https://inference360.ai/badges/mercury-3.svg)](https://inference360.ai/models/mercury-3)

Compare with

Ripple-65B Preview52.3Mercury 2.546.2

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Inception

All models
Mercury 2.546.2Mercury 2.5 PreviewUnrankedMercury Edit 2UnrankedMercury 2UnrankedMercury DecideUnranked