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

GDM-MRCR v2 8-needle, 128K average (GDM-MRCR v2 128K average)

Google's configuration-specific long-context retrieval result at a 128K average context length.

Data verified 28 confirmed releases in the last 30 daysStart free brief

Benchmark score on GDM-MRCR v2 128K average — August 13, 2026

BenchLM mirrors the published score view for GDM-MRCR v2 128K average. Gemini 3.7 Flash leads the public snapshot at 97%. We do not use these results to rank models overall.

1 modelReasoningCurrentDisplay onlyUpdated August 13, 2026

Benchmark score table (1 model)

Score
1
Gemini 3.7 FlashGoogle · Closed
97%

About GDM-MRCR v2 128K average

Year

2026

Tasks

8-needle retrieval tasks

Format

Long-context retrieval

Difficulty

Long-context reasoning

Google reports this 8-needle result at a 128K average context length. We keep it display-only instead of merging it into the weighted MRCRv2 lane.

BenchLM freshness & provenance

Version

GDM-MRCR v2 128K average 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.

FAQ

What does GDM-MRCR v2 128K average measure?

Google's configuration-specific long-context retrieval result at a 128K average context length.

Which model scores highest on GDM-MRCR v2 128K average?

Gemini 3.7 Flash by Google currently leads with a score of 97% on GDM-MRCR v2 128K average.

How many models are evaluated on GDM-MRCR v2 128K average?

1 AI models have been evaluated on GDM-MRCR v2 128K average on BenchLM.

Last updated: August 13, 2026 · BenchLM version GDM-MRCR v2 128K average 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.