A large-scale Scale AI benchmark for tool use with real MCP servers. It covers 36 servers, 220 tools, and 1,000 natural-language tasks that test tool discovery, cross-server orchestration, multi-step calls, and answer synthesis.
Browse the latest scores, model modes, release dates, and parameter sizes for MCP-Atlas.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| Rank | Model | License | |||
|---|---|---|---|---|---|
 Muse Spark 1.2 Thinking EnabledTools | 90.30 | 2026-08-05 | Unknown | Closed | |
Muse Spark 1.1 Thinking EnabledTools | 88.10 | 2026-07-09 | Unknown | Closed | |
 Claude Opus 5 Thinking Level · Extra HighTools | 85.80 | 2026-07-24 | Unknown | Closed | |
4 |  Kimi K3 Thinking Level · MaxTools | 84.20 | 2026-07-16 | 2800B | Conditional |
5 |  Hy4 preview Thinking Level · HighTools | 83.70 | 2026-08-28 | 770B | Free Commercial |
6 |  Gemini 3.5 Flash Thinking Level · HighTools | 83.60 | 2026-06-20 | Unknown | Closed |
7 | Claude Fable 5 Standard ModeTools | 83.30 | 2026-06-09 | Unknown | Closed |
8 | Muse Spark Standard ModeTools | 82.20 | 2026-04-08 | Unknown | Closed |
9 | Claude Opus 4.8 Deep Thinking ModeTools | 82.20 | 2026-05-28 | Unknown | Closed |
10 | Claude Opus 4.8 Thinking Level · MaxTools | 82.20 | 2026-05-28 | Unknown | Closed |
11 | Opus 4.7 Deep Thinking ModeTools | 79.10 | 2026-04-16 | Unknown | Closed |
12 | Opus 4.7 Thinking Level · MaxTools | 79.10 | 2026-04-16 | Unknown | Closed |
13 | Hy3 Thinking Level · HighTools | 79.10 | 2026-07-06 | 295B | Free Commercial |
14 | Gemini 3.1 Pro Preview Thinking Level · HighTools | 78.20 | 2026-02-20 | Unknown | Closed |
15 | Claude Opus 4.6 Deep Thinking ModeTools | 76.80 | 2026-02-05 | Unknown | Closed |
16 | Claude Opus 4.6 Thinking Level · MaxTools | 76.80 | 2026-02-05 | Unknown | Closed |
17 |  GLM-5.2 Thinking EnabledTools | 76.80 | 2026-06-13 | 753.3B | Free Commercial |
18 |  Qwen3.7 Max Thinking EnabledTools | 76.40 | 2026-05-20 | Unknown | Closed |
19 | Kimi K2.7 Code Thinking EnabledTools | 76.00 | 2026-06-12 | 1000B | Free Commercial |
20 | Inkling Thinking Level · Extra HighTools | 76.00 | 2026-07-15 | 975B | Free Commercial |
21 | GLM 5.1 Standard ModeTools | 75.60 | 2026-03-27 | 754B | Free Commercial |
22 | Muse Glimmer-30B Thinking Level · HighTools | 75.50 | 2026-08-10 | 29.6B | Free Commercial |
23 |  GPT-5.5 Thinking Level · Extra HighTools | 75.30 | 2026-04-23 | Unknown | Closed |
24 |  MiniMax M3 Thinking EnabledTools | 74.20 | 2026-06-01 | 428B | Non-Commercial |
25 | Inkling Thinking EnabledTools | 74.10 | 2026-07-15 | 975B | Free Commercial |
26 | GPT-5.4 Thinking Level · Extra HighTools | 70.60 | 2026-03-05 | Unknown | Closed |
27 | Gemini 3.0 Pro (Preview 11-2025) Standard ModeTools | 70.30 | 2025-11-18 | Unknown | Closed |
28 | Opus 4.5 Thinking Level · HighTools | 69.80 | 2025-11-25 | Unknown | Closed |
29 | Claude Sonnet 4.6 Standard ModeTools | 69.50 | 2026-02-17 | Unknown | Closed |
30 | Kimi K2.6 Thinking EnabledTools | 69.40 | 2026-04-20 | 1000B | Free Commercial |
