Benchmark Almanac

LoopsBench

Claude Opus 4.7 25% leads since July 2026
Not in model cards

Agents complete 112 software tasks with dependencies among smaller development units. Executable tests measure resolved tasks and the fraction of units that pass.

The evaluation compares agent models and continuation loops separately. Completed units remain regression checks during later work.

Made by
Microsoft
Released
July 2026
Kind
Long terminal sessions
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Code, Actions
Interaction
Agentic
Environment
Terminal
Refresh
One fixed set
Run by
Anyone
In views
Agents
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 18th, 2026

In the authors’ words

Coding agent infrastructure is shifting from harness engineering toward loop engineering as coding agents are deployed for sustained long-horizon software development. Existing benchmarks often center on localized tasks or end-state outcomes, offering limited insight into sustained execution. We introduce LOOPSBENCH, a long-horizon benchmark for loop engineering in coding agent evaluation. Each task is a dependency DAG over separately testable development units with source-evidenced prerequisite edges. LOOPSBENCH comprises 112 tasks from authentic sources spanning 8 programming languages and 9 domains. Its flow-aware runtime releases tests along the ready frontier and retains completed nodes as regression obligations. We evaluate frontier coding agents paired with widely used loop implementations. The strongest configuration, Opus-4.7 with Claude Code and outer continuation, resolves 25.00% of tasks. Recorded plans recover only part of the source-recovered prerequisite DAG, and regression events remain visible across the evaluated loop profiles. We open source the benchmark data and code, including all tasks, more than 5,300 development units, and executable tests, at microsoft/Loopsbench.

Abstract of LoopsBench: From Harness Engineering to Loop Engineering in Coding Agent Evaluation (arXiv:2608.00267, CC0)
9
Models scored on its own leaderboard

Claude Opus 4.7 leads with 25% among models released by April 2026

Best score per model from the benchmark's own leaderboard, read on November 18th, 2026. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Claude Opus 4.7 continuation loop 25%
  2. 2 GPT-5.5 continuation loop 20.5%
  3. 3 GPT-5.4 continuation loop 18.8%
  4. 4 GLM-5.1 continuation loop 18.8%
  5. 5 DeepSeek-V4-Pro continuation loop 18.8%
  6. 6 Gemini 3.1 Pro continuation loop 14.3%
  7. 7 Qwen 3.6 Plus continuation loop 9.82%
  8. 8 Kimi K2.6 continuation loop 6.25%
  9. 9 Grok 4.1 Fast reasoning, continuation loop 4.46%

Source: Official LoopsBench leaderboard

#ModelScoreSettingDateSource
1 Claude Opus 4.7 25% continuation loop, Claude Code Apr 2026 Official leaderboard
2 GPT-5.5 20.5% continuation loop, Claude Code Apr 2026 Official leaderboard
3 GPT-5.4 18.8% continuation loop, Codex Mar 2026 Official leaderboard
4 GLM-5.1 18.8% continuation loop, Claude Code Apr 2026 Official leaderboard
5 DeepSeek-V4-Pro 18.8% continuation loop, Claude Code Apr 2026 Official leaderboard
6 Gemini 3.1 Pro 14.3% continuation loop, Claude Code Feb 2026 Official leaderboard
7 Qwen 3.6 Plus 9.82% continuation loop, Claude Code Mar 2026 Official leaderboard
8 Kimi K2.6 6.25% continuation loop, Claude Code Apr 2026 Official leaderboard
9 Grok 4.1 Fast 4.46% reasoning, continuation loop, Claude Code Nov 2025 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Anthropic
Own leaderboard Not tested since Jul 2026

Source: Official LoopsBench leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Claude Opus 4.7 GPT-5.4 Gemini 3.1 Pro Grok 4.1 Fast 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Claude Opus 4.7 GPT-5.4

Source: Official LoopsBench leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Terminal agents

All 35 in Terminal agents
  • SkillsBench In use 86-87 containerized expert tasks in 11 domains, run with and without Agent Skills. Verifier tests give pass rate and skill lift. GPT-5.5 66.5%
  • Terminal-Bench Hard In use Agents complete 44 hard tasks from Terminal-Bench Core in the Terminus 2 harness. Artificial Analysis scores pass@1 averaged over 3 repeats. Gemma 4 31B IT 36%
  • Terminal-Bench 4.1 In use The Terminal-Bench 4.0 tasks with hardened verifiers: agents complete 66 hard tasks in containerized terminals across software, ML, science, security, and operations. Test scripts verify each result, scored as resolution rate. Copilot Fusion 71.2%
  • Harbor-Index In use Agents solve about 80 hard tasks drawn from 29 benchmarks run through Harbor adapters, such as HLE, GSO and SWE-bench Pro. Each task keeps its source verifier. GPT-5.5 28.1%
  • Terminal-Bench 2.1 Extended Not in model cards Agents complete long command-line tasks in containers, scored by verification tests. Mercor adds 99 new expert-written tasks to the 89 public ones and reports both scores. Muse Wonder 54%
  • Terminal-Bench Pro Not in model cards Agents solve 400 terminal tasks across eight domains. Harbor distributes the 200 public tasks. Alibaba keeps the other 200 for private evaluation. Claude Opus 4.5 51%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for DeepSeek-V4-Pro. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 9 models we track.
  • Warning Newest score The newest came 7 months ago, in April 2026, for DeepSeek-V4-Pro.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
  • Pass Headroom The best is 25%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on loopsbench.ai: 14 rows for models we track at the last read.
  • Pass Last read November 18th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check