Axis Spatial - Earth-Bench V2: leaderboard

Axis Spatial's Earth-Bench V2 cohort evaluates satellite-product analysis over 117 tasks under methodology 1.0. Every entry preserves its own model effort and agent harness, with one scored attempt per task and a two-hour limit. Code checks grade files and numerical/geometry outputs; Claude Sonnet 5.5 grades maps and traces where needed. Scores use the source's attempted/scored population, including its pending-trace policy; missing attempts are not zeros. These are adapted evaluator runs, not reproductions of the original papers.

Metric: Pass@1 (%). Source: www.axisspatial.com. Saturation forecast: Rough model projection: around 2026. 87 models tracked.

Top models

#ModelScoreOverall rank
1Claude Opus 5.5 (Max) (Axis Spatial Run: native=claude-code-claude-opus-5-5-max; cfg=64f5831c15f78626)80.41
2Claude Opus 5.5 (High) (Axis Spatial Run: native=claude-code-claude-opus-5-5-high; cfg=4142b177c8511092)78.63
3Claude Opus 5.5 (xHigh) (Axis Spatial Run: native=claude-code-claude-opus-5-5-xhigh; cfg=b2f9ca8815f8e318)76.92
4GPT-6 Sol (Medium) (Axis Spatial Run: native=codex-gpt-6-sol-medium; cfg=3ad67e314cdf0a04)76.71
5GPT-6 Sol (High) (Axis Spatial Run: native=codex-gpt-6-sol-high; cfg=5c98741aaf235da5)75.21
6Kimi K3 (Reasoning) (Axis Spatial Run: native=kimi-code-kimi-k3-always-on; cfg=215b566e680a14ae)74.36
7GLM-5.3 (Max) (Axis Spatial Run: native=claude-code-glm-5-3-max-default; cfg=8c7380e9d3eb690c)74.36
8Claude Opus 5 (High) (Axis Spatial Run: native=claude-code-claude-opus-5-high; cfg=beb735020b8f0d3d)73.5
9GPT-6.1 Sol (xHigh) (Axis Spatial Run: native=codex-gpt-6-1-sol-xhigh; cfg=3801b72ee5f0ebfa)71.79
10GPT-6.1 Sol (Medium) (Axis Spatial Run: native=codex-gpt-6-1-sol-medium; cfg=be21af26b13cb8e8)71.79

Interactive version: theaggregate.ai/benchmark?slug=axis-spatial-earth-bench-v2 · How It Works · Data refreshed daily, snapshot 2026-11-05.