DeepPlanning

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

Data verified September 27, 202634 confirmed releases in the last 30 daysFollow model changes

A long-horizon planning benchmark that tests whether agents can optimize under explicit time, budget, and feasibility constraints.

Benchmark score on DeepPlanning — September 27, 2026

We compile the DeepPlanning rows from provider self-reports and secondary reports. Qwen3.7 Plus leads the table at 62.3%, followed by Qwen3.6 Plus (41.5%) and Qwen3.5 397B (37.6%). We do not use these results to rank models overall.

1Proprietary

Qwen3.7 Plus

Alibaba

62.3%

Overall 55.78Context 1M

2Proprietary

Qwen3.6 Plus

Alibaba

41.5%

Overall 54.59Context 1M

3Open weights

Qwen3.5 397B

Alibaba

37.6%

Context 128K

7 modelsAgenticCurrentDisplay onlyUpdated September 27, 2026

Benchmark score table (7 models)

Score

Qwen3.7 PlusAlibaba · Closed

62.3%

Qwen3.6 PlusAlibaba · Closed

41.5%

Qwen3.5 397BAlibaba · Open weight

37.6%

Claude Opus 4.5Anthropic · Closed

26.4%

Qwen3.6-35B-A3BAlibaba · Open weight

25.9%

GLM-5Z.AI · Open weight

14.6%

Kimi K2.5Moonshot AI · Open weight

14.4%

Among the reported DeepPlanning rows, Qwen3.7 Plus is first at 62.3%. The third row is 24.7 points behind. The broader top-10 range is 47.9 points, so the table still separates the published systems.

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

About DeepPlanning

Year

2026

Tasks

Travel planning and constrained shopping

Format

Long-horizon planning benchmark

Difficulty

Constrained agent planning

DeepPlanning focuses on global constrained optimization rather than local next-step reasoning. It is useful because many agents can execute short actions but still fail when they must gather information and plan coherently over a long horizon under hard constraints.

DeepPlanning: Benchmarking Long-Horizon Agentic Planning with Verifiable Constraints

Freshness and provenance

Version

DeepPlanning 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 DeepPlanning measure?

A long-horizon planning benchmark that tests whether agents can optimize under explicit time, budget, and feasibility constraints.

Which model scores highest on DeepPlanning?

Qwen3.7 Plus by Alibaba currently leads with a score of 62.3% on DeepPlanning.

How many models are evaluated on DeepPlanning?

7 AI models have been evaluated on DeepPlanning on BenchLM.

Compare top models on DeepPlanning

Qwen3.7 Plus vs Qwen3.6 Plus Qwen3.6 Plus vs Qwen3.5 397B Qwen3.5 397B vs Claude Opus 4.5 Claude Opus 4.5 vs Qwen3.6-35B-A3B

Last updated: September 27, 2026 · BenchLM version DeepPlanning 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.