Mercury 3
Mercury 3 is capable in instruction following and behind the leaders in reasoning. Too few results yet to rate coding, agentic tasks, safety, long context, math, multimodal tasks, multilingual tasks, or factuality.
Capability profile
Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.
Too few results yet to rate Coding, Agentic, Safety, Long Context, Math, Multimodal, Multilingual or Factuality.
Price
$0.10input$0.35outputper million tokens
From inception · All prices
Evidence
13results on8benchmarks
- 8 independently verified
- 5 vendor-reported
From 4 sources · latest Nov 18, 2026 · How verification works
API features
Tool callingStructured outputsReasoning
As listed by OpenRouter
Mercury 3 benchmark results
13 results on 8 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
20.3% behind the leader2 of 3 ranked benchmarks measured
- 76.84Nov 18, 2026
- 60.35Nov 18, 2026
40.2% behind the leader2 of 6 ranked benchmarks measured
- GPQA Diamond89.23Nov 18, 2026GPQA
- 61.81Nov 17, 2026
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- 80.67Nov 17, 2026
- 84.67Nov 17, 2026
- 88.50Nov 17, 2026
- DeepSearchQA (F1)71.33Nov 18, 2026DeepSearchQA
- τ³-Bench Banking30.24Nov 18, 2026Tau3 Banking
Mercury 3: common questions
Who makes Mercury 3?
Mercury 3 is made by Inception.
What is Mercury 3 good at?
Mercury 3 is capable in instruction following and behind the leaders in reasoning. Too few results yet to rate coding, agentic tasks, safety, long context, math, multimodal tasks, multilingual tasks, or factuality.
How much does Mercury 3 cost?
Mercury 3 costs $0.10 per million input tokens and $0.35 per million output tokens, according to inception. At a mix of three input tokens to one output token, it is cheaper than 83% of the 339 priced models we track.
How many benchmarks has Mercury 3 been tested on?
We track 13 results for Mercury 3 on 8 benchmarks from 4 sources, 8 of them independently verified. The latest was recorded on Nov 18, 2026.
Which API features does Mercury 3 support?
OpenRouter lists tool calling, structured outputs, and reasoning for Mercury 3.
About this record
Where Mercury 3's numbers come from, and every name it appears under.
- Tracked since
- Oct 19, 2026
- Newest source mention
- Nov 17, 2026
Where the results come from
Verification: 13 scores · 8 independently verified · 5 vendor-reported. How these tiers are assigned
From 4 sources on 2 sites. arcprize.org supplies 8 of them; the 8 independently verified results come from 1 site. Bars are coloured by trust tier.
- arcprize.org8
- api.llm-stats.com5
Also known as
How our sources name Mercury 3 at each reasoning setting.
| Setting | Short form |
|---|---|
| low | mercury 3 preview (low) |
| medium | mercury 3 preview (medium) |
| high | mercury 3 preview (high) |