Skip to main content
Radar

Keep up with the models you depend on. Follow price changes, retirements, and API updates.Follow the models you depend on.

Follow model changes
BenchLM

MCP Atlas

A benchmark for tool-calling over Model Context Protocol integrations and external tools.

Data verified 31 confirmed releases in the last 30 daysSee provider release alerts

Benchmark score on MCP Atlas — September 10, 2026

We mirror the published score view for MCP Atlas. Muse Spark 1.1 leads the public snapshot at 88.1%, followed by Claude Opus 5 (85.8%) and Kimi K3 (84.2%). We do not use these results to rank models overall.

1Proprietary

Muse Spark 1.1

Meta

88.1%
Overall 70.41Context 1M
2Proprietary

Claude Opus 5

Anthropic

85.8%
Overall 81.97
3Proprietary

Kimi K3

Moonshot AI

84.2%
Overall 74.82Context 1.05M

41 modelsAgenticCurrentDisplay onlyUpdated September 10, 2026

Benchmark score table (41 models)

Score
1
Muse Spark 1.1Meta · Closed
88.1%
2
Claude Opus 5Anthropic · Closed
85.8%
3
Kimi K3Moonshot AI · Closed
84.2%
4
Hy4 previewTencent · Open weight
83.7%
5
Gemini 3.5 FlashGoogle · Closed
83.6%
6
Claude Opus 4.8Anthropic · Closed
82.2%
7
Ornith-1.5-397BOrnith AI · Open weight
80%
8
Inkling-SmallThinking Machines Lab · Open weight
79.6%
9
Claude Opus 4.7 (Adaptive)Anthropic · Closed
77.3%
10
GLM-5.2Z.AI · Open weight
76.8%
11
Qwen3.7 MaxAlibaba · Closed
76.4%
12
Kimi K2.7 CodeMoonshot AI · Open weight
76%
13
Muse Glimmer 30BMeta · Open weight
75.5%
14
GPT-5.5OpenAI · Closed
75.3%
15
MiniMax M3MiniMax · Open weight
74.2%
16
DeepSeek V4 Pro (High)DeepSeek · Open weight
74.2%
17
InklingThinking Machines Lab · Open weight
74.1%
18
DeepSeek V4 Pro 0813DeepSeek · Closed
73.6%
19
Qwen3.7 PlusAlibaba · Closed
73.2%
20
GLM-5.1Z.AI · Open weight
71.8%
21
GPT-5.4OpenAI · Closed
70.6%
22
Ornith-1.5-35B-A3BOrnith AI · Open weight
70.2%
23
DeepSeek V4 ProDeepSeek · Open weight
69.4%
24
DeepSeek V4 Flash 0731DeepSeek · Closed
69%
25
DeepSeek V4 Flash (High)DeepSeek · Closed
67.4%
26
Ling 3.0 FlashInclusionAI · Open weight
65.5%
27
DeepSeek V4 FlashDeepSeek · Closed
64%
28
Qwen3.6-35B-A3BAlibaba · Open weight
62.8%
29
Solar Pro 4Upstage · Closed
61.4%
30
Solar Open 2Upstage · Open weight
58.2%
31
GPT-5.4 miniOpenAI · Closed
57.7%
32
GPT-5.4 nanoOpenAI · Closed
56.1%
33
Kimi K2.6Moonshot AI · Open weight
55.9%
34
Ornith-1.5-9BOrnith AI · Open weight
54.2%
35
Qwen3.6 PlusAlibaba · Closed
48.2%
36
LLaDA2.2-flashInclusionAI · Open weight
46.2%
37
Qwen3.5 397BAlibaba · Open weight
46.1%
38
LongCat-Flash-Lite-SparseMeituan · Open weight
45.6%
39
Claude Opus 4.5Anthropic · Closed
42.3%
40
GLM-5Z.AI · Open weight
31.1%
41
Kimi K2.5Moonshot AI · Open weight
29.5%

The published MCP Atlas snapshot places Muse Spark 1.1 first at 88.1%. The third row is 3.9 points behind. The broader top-10 range is 11.3 points, so the table still separates the published systems.

41 models have been evaluated on MCP Atlas. The benchmark falls in the Agentic category. This category carries a 22% weight in BenchLM.ai's overall scoring system. MCP Atlas is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MCP Atlas

Year

2026

Tasks

Tool-integrated agent tasks

Format

Interactive tool-calling evaluation

Difficulty

Advanced tool use

OpenAI reports MCP Atlas as a tool-use benchmark that measures how well models work with MCP-backed systems and external tools.

Introducing GPT-5.4 mini and nano

BenchLM freshness & provenance

Version

MCP Atlas 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does MCP Atlas measure?

A benchmark for tool-calling over Model Context Protocol integrations and external tools.

Which model scores highest on MCP Atlas?

Muse Spark 1.1 by Meta currently leads with a score of 88.1% on MCP Atlas.

How many models are evaluated on MCP Atlas?

41 AI models have been evaluated on MCP Atlas on BenchLM.

Compare Top Models on MCP Atlas

Muse Spark 1.1 vs Claude Opus 5Claude Opus 5 vs Kimi K3Kimi K3 vs Hy4 previewHy4 preview vs Gemini 3.5 Flash
Last updated: September 10, 2026 · BenchLM version MCP Atlas 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.