Skip to main content
Radar

Five or fewer confirmed AI changes, with original sources, on mornings when something changed.A free source-linked morning brief.

Start free brief
BenchLM

LongBench v2

A long-context benchmark that measures whether models can actually use extended context windows for reasoning and retrieval.

Data verified 29 confirmed releases in the last 30 daysStart free brief

Top models on LongBench v2 — August 19, 2026

As of August 19, 2026, Qwen3.8 Max leads the LongBench v2 leaderboard with 66.3% , followed by Claude Opus 4.5 (64.4%) and Qwen3.5 397B (63.2%).

1Open

Qwen3.8 Max

Alibaba

66.3%
Overall 79.94Context 1M
2Closed

Claude Opus 4.5

Anthropic

64.4%
Overall 63.65Context 200K
3Open

Qwen3.5 397B

Alibaba

63.2%
Overall 57.44Context 128K
13 modelsReasoning38% of category scoreCurrentUpdated August 19, 2026

Leaderboard (13 models)

Score
1
Qwen3.8 MaxAlibaba · Open weight
66.3%
2
Claude Opus 4.5Anthropic · Closed
64.4%
3
Qwen3.5 397BAlibaba · Open weight
63.2%
4
Qwen3.6 PlusAlibaba · Closed
62%
5
Nemotron 3 UltraNVIDIA · Open weight
61.9%
6
Kimi K2.5Moonshot AI · Open weight
61%
7
GLM-5Z.AI · Open weight
60.8%
8
Qwen3.5-27BAlibaba · Open weight
60.6%
9
Qwen3.5-122B-A10BAlibaba · Open weight
60.2%
10
Agents-A1InternScience · Open weight
60.2%
11
Qwen3.5-35B-A3BAlibaba · Open weight
59%
12
DeepSeek V4 Pro BaseDeepSeek · Open weight
51.5%
13
DeepSeek V4 Flash BaseDeepSeek · Open weight
44.7%

According to BenchLM.ai, Qwen3.8 Max leads the LongBench v2 benchmark with a score of 66.3%, followed by Claude Opus 4.5 (64.4%) and Qwen3.5 397B (63.2%). The scores show moderate spread, with meaningful differences between the top tier and mid-tier models.

13 models have been evaluated on LongBench v2. The benchmark falls in the Reasoning category. This category carries a 17% weight in BenchLM.ai's overall scoring system. Within that category, LongBench v2 contributes 38% of the category score, so strong performance here directly affects a model's overall ranking.

About LongBench v2

Year

2025

Tasks

Long-context tasks

Format

Extended-context retrieval and reasoning

Difficulty

Hard long-context

LongBench v2 is useful because context-window size alone is not a capability. It measures whether a model can retain, retrieve, and reason over long inputs effectively.

LongBench v2

BenchLM freshness & provenance

Version

LongBench v2 2025

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

Current

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 LongBench v2 measure?

A long-context benchmark that measures whether models can actually use extended context windows for reasoning and retrieval.

Which model scores highest on LongBench v2?

Qwen3.8 Max by Alibaba currently leads with a score of 66.3% on LongBench v2.

How many models are evaluated on LongBench v2?

13 AI models have been evaluated on LongBench v2 on BenchLM.

Compare Top Models on LongBench v2

Qwen3.8 Max vs Claude Opus 4.5Claude Opus 4.5 vs Qwen3.5 397BQwen3.5 397B vs Qwen3.6 PlusQwen3.6 Plus vs Nemotron 3 Ultra
Last updated: August 19, 2026 · BenchLM version LongBench v2 2025

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.