Every AI leaderboard

Leaderboards28Models ranked986

28 leaderboards · 986 models ranked

MCP Atlas leaderboard: who is #1

Scale AI · Pass rate, higher is better · 37 models · published Nov 9, 2026

Scale AI

Top 25 of 37 models, best variant per model as published

#ModelScoreCorrect
1New, entered at #1 since Nov 6, 2026Gemini 4 Argonas “Gemini-4-Argon”90.7%
21down from #1 since Nov 6, 2026Muse Spark 1.1Meta88.1%
31down from #2 since Nov 6, 2026Claude Fable 5.1as “Fable 5.1”87.2%
41down from #3 since Nov 6, 2026Claude Opus 5as “claude-opus-5 (xhigh)”85.8%
5New, entered at #5 since Nov 6, 2026Claude Opus 5.5as “Opus 5.5”85.4%
5New, entered at #5 since Nov 6, 2026Inkling Small 285.4%
73down from #4 since Nov 6, 2026Qwen3.8 2.4T A95Bas “Qwen3.8-2.4T-A95B (xHigh)”84.5%
83down from #5 since Nov 6, 2026GLM-5.3as “GLM 5.3”84.2%
93down from #6 since Nov 6, 2026Gemini 3.5 Flashas “Gemini 3.5 Flash (high)”83.6%
103down from #7 since Nov 6, 2026Claude Fable 5Anthropic83.3%
113down from #8 since Nov 6, 2026Kimi K3as “kimi-k3 (max)”82.3%
123down from #9 since Nov 6, 2026Claude Opus 4.8as “claude-opus-4-8 (max)”82.2%
123down from #9 since Nov 6, 2026Muse SparkMeta82.2%
143down from #11 since Nov 6, 2026GPT-5.6as “gpt-5.6 (sol)”81.8%
153down from #12 since Nov 6, 2026Inkling Smallas “Inkling-small”79.2%
163down from #13 since Nov 6, 2026Claude Opus 4.7as “claude-opus-4-7 (max)”79.1%
173down from #14 since Nov 6, 2026Gemini 3.1 Pro Previewas “gemini-3.1-pro-preview (high)”78.2%
183down from #15 since Nov 6, 2026GLM-5p2as “glm-5p2”77.8%
193down from #16 since Nov 6, 2026Claude Opus 4.6as “claude-opus-4-6 (max)”76.8%
203down from #17 since Nov 6, 2026Inklingas “Inkling (xHigh)”76.0%
213down from #18 since Nov 6, 2026GLM-5p1as “glm-5p1”75.6%
223down from #19 since Nov 6, 2026GPT-5.5as “gpt-5.5 (xhigh)”75.3%
233down from #20 since Nov 6, 2026GPT-5.4as “gpt-5.4 (xhigh)”70.6%
243down from #21 since Nov 6, 2026Gemini 3 Proas “gemini-3-pro-preview”70.3%
253down from #22 since Nov 6, 2026Claude Opus 4.5as “claude-opus-4-5 (high)”69.8%
About this board and its history

What it measures

Real tasks over 36 MCP servers and 220 tools, three to six tool calls each (the 500-task public subset).

Rank at each month's end

12345678910SepOctNow1Gemini 4 Argon2Muse Spark 1.13Claude Fable 5.14Claude Opus 55Claude Opus 5.55Inkling Small 2
Gemini 4 ArgonMuse Spark 1.1Claude Fable 5.1Claude Opus 5Claude Opus 5.5Inkling Small 2

Other agents and tools boards

Source

Scores as published by Scale AI on the capture date (2026-11-09). Source: labs.scale.com leaderboard page.

labs.scale.com/leaderboard/mcp_atlas