AA Coding Agent Index v2.1 - Terminal-Bench 4.1: leaderboard

Artificial Analysis Coding Agent Index v2.1 (November 2026) evaluates a model inside its named coding-agent harness. It averages FrontierCode-AA 2.0 (132 repository-editing tasks), Terminal-Bench 4.1 (66 terminal tasks with revised verifiers), and SWE-Atlas-QnA-AA (124 repository questions), each using three attempts per task and task-equal means. Audited reward-hacking attempts score zero; environment/verifier errors are excluded. Native model effort and harness remain separate; this edition differs from v2.0. This board reports Terminal-Bench 4.1, higher is better.

Metric: Pass@1 (%). Source: artificialanalysis.ai. Saturation forecast: Around April 2027. 23 models tracked. Updated 14 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Kimi K3.1 Pro67.17#29
2Claude Sonnet 5.5Max, Default Fallback, AA Coding Agent Run: harness=Claude Code; native=Claude Sonnet 5.5 (Max, Default Fallback); fallback=source-published65.66#19
3Gemini 4 CarbonHigh64.65#2
4Claude Opus 5.5Max, Default Fallback, AA Coding Agent Run: harness=Claude Code; native=Claude Opus 5.5 (Max, Default Fallback); fallback=source-published64.65#5
5Claude Fable 5.2Xhigh, Default Fallback, AA Coding Agent Run: harness=Claude Code; native=Claude Fable 5.2 (Xhigh, Default Fallback); fallback=source-published64.14#1
6Muse Wonder FullMax59.6#17
7Claude Fable 5.1Max, Default Fallback, AA Coding Agent Run: harness=Claude Code; native=Claude Fable 5.1 (Max, Default Fallback); fallback=source-published58.08#15
8GPT-6 AstraMax57.58#6
9Gemini 4 ArgonHigh56.57#12
10GPT-6.1 SolMax54.04#10
11GPT-6.1 TerraMax50.51#21
12Muse WonderMax47.47#28
13Muse EmberMax43.43#53
14Mistral Large 4High42.42#137
15GLM-5.3Max40.4#68

Interactive version: theaggregate.ai/benchmark?slug=aa-coding-agent-index-v2-1-terminal-bench-4-1 · How It Works · Data refreshed daily, snapshot 2026-11-16.