Skip to main content
BenchLM
Data

Artificial Analysis Coding Agent Index (AA Coding Agents)

We show this table for reference; we do not rank on it.

A display-only Artificial Analysis leaderboard for coding-agent systems, combining agent harnesses, host models, and execution settings across software-engineering benchmarks.

Index score on AA Coding Agents — Source captured 2026-11-09; evaluation dates not inferred

We mirror the published index score view for AA Coding Agents. Claude Code - Claude Opus 5.5 (Max, Default Fallback) leads the public snapshot at 61.7%, followed by Devin Fusion CLI - Claude Opus 5.5 Max + GPT-6 Luna High (61.5%) and GitHub Copilot - Copilot Fusion (max) (59.9%). We do not use these results to rank models overall.

1

Claude Code - Claude Opus 5.5 (Max, Default Fallback)

Claude Code

61.7%
2

Devin Fusion CLI - Claude Opus 5.5 Max + GPT-6 Luna High

Devin Fusion CLI

61.5%
3

GitHub Copilot - Copilot Fusion (max)

GitHub Copilot

59.9%

35 modelsCodingCurrentDisplay onlyUpdated Source captured 2026-11-09; evaluation dates not inferred

Index score table (35 models)

Index score results for AA Coding Agents
RankModel / configurationScoreParameters (B)Open / closed
1Claude Code - Claude Opus 5.5 (Max, Default Fallback)Claude Code
61.7%
Not reportedUnknown
2Devin Fusion CLI - Claude Opus 5.5 Max + GPT-6 Luna HighDevin Fusion CLI
61.5%
Not reportedUnknown
3GitHub Copilot - Copilot Fusion (max)GitHub Copilot
59.9%
Not reportedUnknown
4Antigravity CLI - Gemini 4 Carbon (High)Antigravity CLI
59.4%
Not reportedUnknown
5Claude Code - Claude Sonnet 5.5 (Max, Default Fallback)Claude Code
59.3%
Not reportedUnknown
6Muse Code - Muse Wonder Full (Max)Muse Code
58.9%
Not reportedUnknown
7Devin Fusion CLI - Claude Fable 5.1 XHigh + SWE-2 MediumDevin Fusion CLI
58.7%
Not reportedUnknown
8Claude Code - Claude Fable 5.1 (Max, Default Fallback)Claude Code
57.6%
Not reportedUnknown
9Codex - GPT-6 Astra (Max)Codex
54.7%
Not reportedUnknown
10Codex - GPT-6.1 Sol (Max)Codex
54.1%
Not reportedUnknown
11Antigravity CLI - Gemini 4 Argon (High)Antigravity CLI
53.2%
Not reportedUnknown
12Devin Fusion CLI - GPT-6 Astra XHigh + SWE-2 MediumDevin Fusion CLI
53.1%
Not reportedUnknown
13Muse Code - Muse Wonder (Max)Muse Code
52.0%
Not reportedUnknown
14Opencode - GLM-5.3 (Max)Opencode
48.9%
Not reportedUnknown
15Codex - GPT-6.1 Terra (Xhigh)Codex
48.2%
Not reportedUnknown
16Codex - GPT-6.1 Terra (Max)Codex
48.1%
Not reportedUnknown
17Codex - DeepSeek V4.1 Flash (Max)Codex
47.3%
Not reportedUnknown
18Grok Build - Grok 4.7 (Xhigh)Grok Build
46.6%
Not reportedUnknown
19DeepSeek Harness - DeepSeek V4.1 Flash 1017 (Max)DeepSeek Harness
46.3%
Not reportedUnknown
20Codex - GPT-6.1 Terra (High)Codex
46.3%
Not reportedUnknown
21Muse Code - Muse Ember (Max)Muse Code
45.9%
Not reportedUnknown
22Muse Code - Muse Spark 1.3 (Max)Muse Code
45.7%
Not reportedUnknown
23Claude Code - Step 5 PreviewClaude Code
43.1%
Not reportedUnknown
24Codex - GPT-6.1 Terra (Medium)Codex
42.0%
Not reportedUnknown
25Kimi Code CLI - Kimi K3 (Max)Kimi Code CLI
41.4%
Not reportedUnknown
26Muse Code - Muse Ember (Xhigh)Muse Code
39.8%
Not reportedUnknown
27Claude Code - Qwen3.8 MaxClaude Code
39.2%
Not reportedUnknown
28Codex - GPT-6.1 Terra (Non-reasoning)Codex
39.2%
Not reportedUnknown
29Claude Code - Claude Haiku 5.5 (Max)Claude Code
37.3%
Not reportedUnknown
30Mistral Vibe - Mistral Large 4 (High)Mistral Vibe
35.4%
Not reportedUnknown
31Codex - GPT-6.1 Terra (Low)Codex
34.9%
Not reportedUnknown
32Antigravity SDK - Gemini 3.8 Flash (High)Antigravity SDK v0.1.12
33.3%
Not reportedUnknown
33Codex - GPT-6 Luna (Max)Codex
31.2%
Not reportedUnknown
34Muse Code - Muse Spark Open (Xhigh)Muse Code
26.3%
Not reportedUnknown
35Mirror - Beam 1.0Mirror
20.5%
Not reportedUnknown

How AA Coding Agents is shown here

BenchLM mirrors the Artificial Analysis Coding Agent Index v1.1 page as a display-only agent leaderboard. The source compares coding-agent variants across FrontierCode v1.1 Main, SWE-Atlas-QnA, Terminal-Bench v4 and reports the average pass@1 index alongside cost, token, and execution-time metadata.

AA Coding Agents is separate from BenchLM model-only rankings. Its rows combine an agent harness, a host model, execution settings, and provider routing, so BenchLM treats the index as external system evidence rather than a weighted base-model benchmark. Component benchmark availability can vary by row in the source payload.

Snapshot

35 indexed rows35 source models3 component benchmarksv2.0Cost/time/token metadataDisplay only
Artificial Analysis coding agentsArtificial Analysis methodologyDeepSWE componentTerminal-Bench component

The published AA Coding Agents snapshot places Claude Code - Claude Opus 5.5 (Max, Default Fallback) first at 61.7%. The third row is 1.8 points behind. The broader top-10 range is 7.6 points, so many of the published results sit in a relatively narrow band.

35 models have been evaluated on AA Coding Agents. The benchmark falls in the Coding category. AA Coding Agents is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About AA Coding Agents

Year

2026

Tasks

Composite over DeepSWE, Terminal-Bench v2, and SWE-Atlas-QnA

Format

Average pass@1 index

Difficulty

Real-world coding-agent workflows

BenchLM mirrors the Artificial Analysis Coding Agent Index v1.1 page as a display-only external leaderboard. The source combines DeepSWE, Terminal-Bench v2, and SWE-Atlas-QnA component scores and publishes cost, token, and execution-time metadata. Rows are coding-agent systems rather than pure base-model results.

Artificial Analysis Coding Agent BenchmarksPublic benchmark source

Freshness and provenance

Version

AA Coding Agents 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.

Questions

What does AA Coding Agents measure?

A display-only Artificial Analysis leaderboard for coding-agent systems, combining agent harnesses, host models, and execution settings across software-engineering benchmarks.

Which model leads the published AA Coding Agents snapshot?

Claude Code - Claude Opus 5.5 (Max, Default Fallback) currently leads the published AA Coding Agents snapshot with 61.7% index score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on AA Coding Agents?

The Source captured 2026-11-09; evaluation dates not inferred snapshot contains 35 AI models.

Last updated: Source captured 2026-11-09; evaluation dates not inferred · mirrored from the public benchmark leaderboard

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 6,100+ subscribers.

One email each week. Unsubscribe anytime.