whatmodel.sh / decision / Mercury Decide (paid endpoint)
Mercury Decide (paid endpoint)
↗ share
+ contribute
Mercury Decide (paid endpoint) is a closed-weights decision model from Inception. Licence: proprietary.
Catalogue identity notes
Named hosted endpoint verified 2026-10-31. Perch records the requested model and test date; the service may change its weights. No immutable current-weight equivalence asserted. Paid endpoint inception/mercury-decide, recorded revision inception/mercury-decide-20260930; kept separate from the free endpoint.
$ how it compares
The decision models ranked either side of Mercury Decide (paid endpoint), out of 357 with a score, then more from Inception.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 54 |  StartLux-Decision v1.1 4B | 65±1 | model files | ||
| 55 | Quyet-1.0-Large | 65±1 | model files | ||
| 56 | Mercury Decide (paid endpoint) | 65±1 | api price unknown | ||
| 57 |  Jebadiah 27B v2.1 (T1e) | 65±1 | local | ||
| 58 |  OneJev1.5-27B | 65±4 | model files | ||
| more from Inception | |||||
| 77 | Mercury Decide (free endpoint) | 63±2 | api price unknown | ||
see all decision models · find the ones that fit your machine
$ what it's good at
retrieval & classification▁▁▁·▁▁▁▁▃▃█▅ #83 of 258 * · 6.1% of its score · 2 tests
Beats AutoTrust-JEV-27B v1.1, oenogarum (Gemma 4 31B), Decider 31B (NVFP4) here, though they rank higher overall. Trails StartLux-Decision v1.1 0.8B, Standard One 3B SH (v4), ezjev 4B s2 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LeanMCP · scifact rank noul | 74.3% ✎ | ▂····▂···▃▅█ beats 8/11 | 3.2% |
| LeanMCP · nfcorpus rank noul | 37.5% ✎ | ▃▃·····▃▃▅██ beats 8/11 | 2.9% |
knowledge & reasoning▁▂▂▂▂▃▃▅▅▆█▂ #36 of 263 · 48% of its score · 7 tests
Beats vera-large, KnowLine 26B-A4B Gen2.5, OneJev1.6-27B here, though they rank higher overall. Trails Solar Decide, Clef-omni, GPT-6 Luna Decisions here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Etsabary · 1342 text decisions | 75.8% ✎ | ▁··▂·▃▅▁▃▂▂█ beats 12/24 | 19% |
| LeanMCP · scienceqa text | 94.2% ✎ | ▄··▄█▄█▄·█▄▄ beats 7/11 | 6.3% |
| LeanMCP · medqa usmle | 84.9% ✎ | ▄▄▄·▄▄▄▄▄·██ beats 7/11 | 6.1% |
| MedQA · Freyer et al. (first-valid accuracy) | 85.2% ✎ | █··▃▃·▃▃▃·▃█ beats 7/11 | 4.7% |
| LeanMCP · pubmedqa | 65.6% ✎ | ▂···▂·▂·▄▄▂█ beats 4/11 | 4.7% |
| LeanMCP · medmcqa | 72.6% ✎ | ▃·▃··▃█▃·▅▅▃ beats 6/11 | 4.5% |
| LeanMCP · mmlu pro | 64.5% ✎ | ▄··▄▄████··▄ beats 3/11 | 3.5% |
vision▁·▃▂▃▃▇▄▄█▄▃ #9 of 79 * · 13% of its score · 1 test
Beats Fastino GLiDE 1.2, StartLux-Decision v1.1 27B, vLLM-SR d3-ultra (Decision 3.0, 27B) here, though they rank higher overall. Trails Mercury Decide (free endpoint), Clef, Jev 1.14 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Etsabary · 200 image decisions | 71.0% ✎ | ▂▂··▂▃▂·█▃▂▂ beats 7/14 | 13% |
workflows▁▁▁▂▁▂▂▆█▄▃▁ #3 of 114 · 31% of its score · 22 tests
Beats baselabs mammoth-1 (27B), Nace.AI Drex Max v1 (27B), Perplexity Decider v1.2 (27B) here, though they rank higher overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LeanMCP · mcpbench server routing | 100.0% ✎ | █·███··▄···▄ beats 9/9 | 5.7% |
| LeanMCP · mcpmark server routing | 57.1% ✎ | ▂··▂··▄█▂·▂▄ beats 10/11 | 4.5% |
| LeanMCP · rewardbench2 | 79.3% ✎ | ▄▄▄·▄▄▄▄█·█▄ beats 10/11 | 3.9% |
| LeanMCP · aegis2 response | 77.2% ✎ | █··▃·▃·█··▅▅ beats 4/11 | 3.3% |
| LeanMCP · helpsteer2 helpfulness | 0.9469 ordinal levels ✎ | ▄▄▄▄▄·▄·█▄▄▄ beats 7/10 | 2.6% |
| LeanMCP · sst5 | 0.4768 ordinal levels ✎ | ▂·····▂▂▄█▄▃ beats 11/11 | 2.5% |
| LeanMCP · atbench500 | 68.8% ✎ | ▃··▅▅█··▅··▅ beats 2/11 | 2.2% |
| LeanMCP · mcpbench tool selection | 100.0% ✎ | ▂··········█ beats 2/9 | 2.1% |
14 more tests
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LeanMCP · prompt injections | 81.0% ✎ | ▂▂▂▂█·▄▃···▂ beats 10/11 | 1.7% |
| LeanMCP · mcpmark tool selection | 100.0% ✎ | ▂··········█ beats 2/11 | 1.1% |
| Perch semantic lint · private (Youden J) | 0.7887 J ✎ | █····▄█▄···█ beats 7/7 | 0.6% |
| Perch semantic lint · public (Youden J) | 0.7523 J ✎ | █····▄█▄···█ beats 7/7 | 0.6% |
| Perch security detection · public (Youden J) | 0.0929 J ✎ | █···▄▄···▄▄█ beats 4/7 | 0.1% |
| Perch bug detection · private (Youden J) | 0.0436 J ✎ | ▅··▃▃█·····▃ beats 4/7 | 0.1% |
| Perch bug detection · public (Youden J) | 0.0318 J ✎ | █▄·█·▄▄····▄ beats 4/7 | 0.1% |
| Perch security detection · private (Youden J) | 0.0210 J ✎ | █▃▃··▃·▃···▃ beats 4/7 | 0.1% |
| LeanMCP · jailbreak classification | 96.8% ✎ | ▂·▄··▂···█▆▂ beats 6/11 | 0.1% |
| Prism-1 study · Banking20 | 91.3% ✎ | █··········█ beats 1/1 | — |
| Prism-1 study · Newsgroups | 73.3% ✎ | █··········█ beats 0/1 | — |
| Prism-1 study · BEV decision mix | 75.1% ✎ | █··········█ beats 0/1 | — |
| Prism-1 study · Typed decisions | 67.3% ✎ | █··········█ beats 0/1 | — |
| Prism-1 study · JevBench held-out | 89.9% ✎ | █··········█ beats 1/1 | — |
$ where to run it
Through an API
api · OpenRouter · price unknown · openrouter.ai ↗
Model repository and documentation: openrouter.ai ↗
$ how the score was worked out · LeanMCP · JevBench 61%, Etsabary · decision benchmarks 32%, Freyer et al. · MedQA 5% and 1 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 LeanMCP · JevBench | 18 | 61% |
| Etsabary · decision benchmarks | 2 | 32% |
| Freyer et al. · MedQA | 1 | 5% |
 Perch benchmark results | 6 | 2% |
| prior · pull toward the middle | 1% | |
| weighted average | = 65.2 |
not counted (5)
| Prism-1: Position- and Rival-Independent Scoring for System One Models | Prism-1 study · Banking20 [correct] | 91.3% | fewer than three models have this test yet, or this source ran it for this model alone |
| Prism-1: Position- and Rival-Independent Scoring for System One Models | Prism-1 study · Newsgroups [correct] | 73.3% | fewer than three models have this test yet, or this source ran it for this model alone |
| Prism-1: Position- and Rival-Independent Scoring for System One Models | Prism-1 study · BEV decision mix [correct] | 75.1% | fewer than three models have this test yet, or this source ran it for this model alone |
| Prism-1: Position- and Rival-Independent Scoring for System One Models | Prism-1 study · Typed decisions [correct] | 67.3% | fewer than three models have this test yet, or this source ran it for this model alone |
| Prism-1: Position- and Rival-Independent Scoring for System One Models | Prism-1 study · JevBench held-out [correct] | 89.9% | fewer than three models have this test yet, or this source ran it for this model alone |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json