<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en-GB" xml:base="https://www.axisspatial.com/"><title>Geospatial Agent Benchmarks: updates</title><link href="https://www.axisspatial.com/arena/updates.xml" rel="self"/><link href="https://www.axisspatial.com/arena/updates"/><id>https://www.axisspatial.com/arena/updates.xml</id><updated>2026-10-30T00:00:00Z</updated><author><name>Axis Spatial</name></author><entry><title>Release 1.3</title><link href="https://www.axisspatial.com/arena/updates#2026-10-30-release-1-3"/><id>https://www.axisspatial.com/arena/updates#2026-10-30-release-1-3</id><updated>2026-10-30T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;A page for each lab: every model we benchmark from it, on every evaluation.&lt;/li&gt;&lt;li&gt;The home page is now an overview, and the Geospatial Intelligence Index has its own page.&lt;/li&gt;&lt;li&gt;Comparison pages open with the answer, and every page has a model search.&lt;/li&gt;&lt;li&gt;A simpler phone header with a grouped menu.&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>Nemotron 3.5 Super VL 120B: Geospatial knowledge score corrected</title><link href="https://www.axisspatial.com/arena/updates#2026-10-28-nemotron-3-5-super-vl-120b-geospatial-knowledge-score-corrected"/><id>https://www.axisspatial.com/arena/updates#2026-10-28-nemotron-3-5-super-vl-120b-geospatial-knowledge-score-corrected</id><updated>2026-10-28T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;Its Geospatial knowledge score is 92, not 89. The provider returned empty replies to some questions, which had been graded as wrong; they were asked again.&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>Release 1.2</title><link href="https://www.axisspatial.com/arena/updates#2026-10-28-release-1-2"/><id>https://www.axisspatial.com/arena/updates#2026-10-28-release-1-2</id><updated>2026-10-28T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;A small chart for each evaluation on every model page.&lt;/li&gt;&lt;li&gt;Comparisons open on Anthropic against OpenAI, tier for tier, and follow each family&amp;#x27;s newest version.&lt;/li&gt;&lt;li&gt;Safety and Geospatial knowledge results are given as rates.&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>Release 1.1</title><link href="https://www.axisspatial.com/arena/updates#2026-10-27-release-1-1"/><id>https://www.axisspatial.com/arena/updates#2026-10-27-release-1-1</id><updated>2026-10-27T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;A Tasks page with a world map of where the tasks are set.&lt;/li&gt;&lt;li&gt;A Safety page: how often each entry tried to game the checks.&lt;/li&gt;&lt;li&gt;Pages for six more evaluations, among them OGD4All, GeoChrono and BigEarthNet.txt.&lt;/li&gt;&lt;li&gt;Every developer and every model family on every chart.&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>First results for Claude Opus 5.5 (low), Claude Sonnet 5.5 (low) and Claude Haiku 5.5 (low)</title><link href="https://www.axisspatial.com/arena/updates#2026-10-25-first-results-for-claude-opus-5-5-low-claude-sonnet-5-5-low-and-claude-haiku-5-5-low"/><id>https://www.axisspatial.com/arena/updates#2026-10-25-first-results-for-claude-opus-5-5-low-claude-sonnet-5-5-low-and-claude-haiku-5-5-low</id><updated>2026-10-25T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/claude-code-claude-opus-5-5-low&quot;&gt;Claude Opus 5.5 (low), Claude Code&lt;/a&gt;: Geospatial Agent Index 74&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/claude-code-claude-sonnet-5-5-low&quot;&gt;Claude Sonnet 5.5 (low), Claude Code&lt;/a&gt;: Geospatial Agent Index 69&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/claude-code-claude-haiku-5-5-low&quot;&gt;Claude Haiku 5.5 (low), Claude Code&lt;/a&gt;: Geospatial Agent Index 70&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>First results for MiniMax M3.1 Flash (max), Nemotron 3 Nano Omni 30B (Reasoning) and 3 more</title><link href="https://www.axisspatial.com/arena/updates#2026-10-24-first-results-for-minimax-m3-1-flash-max-nemotron-3-nano-omni-30b-reasoning-and-3-more"/><id>https://www.axisspatial.com/arena/updates#2026-10-24-first-results-for-minimax-m3-1-flash-max-nemotron-3-nano-omni-30b-reasoning-and-3-more</id><updated>2026-10-24T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/minimax-m3-1-flash-max&quot;&gt;MiniMax M3.1 Flash (max), Terminus-2&lt;/a&gt;: Geospatial Agent Index 52&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/nemotron-3-nano-omni-30b-reasoning&quot;&gt;Nemotron 3 Nano Omni 30B (Reasoning), Terminus-2&lt;/a&gt;: Geospatial Agent Index 21&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/nemotron-3-5-lightning-30b-reasoning&quot;&gt;Nemotron 3.5 Lightning 30B (Reasoning), Terminus-2&lt;/a&gt;: Geospatial Agent Index 13&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/nemotron-3-5-super-vl-120b-reasoning&quot;&gt;Nemotron 3.5 Super VL 120B (Reasoning), Terminus-2&lt;/a&gt;: Geospatial Agent Index 52&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/step-5-preview-reasoning&quot;&gt;Step 5 Preview (Reasoning), Terminus-2&lt;/a&gt;: Geospatial Agent Index 40&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>First results for Nemotron 3 Ultra (high), DeepSeek V4.1 Flash (max) and Qwen 3.8 27B (xhigh)</title><link href="https://www.axisspatial.com/arena/updates#2026-10-23-first-results-for-nemotron-3-ultra-high-deepseek-v4-1-flash-max-and-qwen-3-8-27b-xhigh"/><id>https://www.axisspatial.com/arena/updates#2026-10-23-first-results-for-nemotron-3-ultra-high-deepseek-v4-1-flash-max-and-qwen-3-8-27b-xhigh</id><updated>2026-10-23T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/nemotron-3-ultra-high&quot;&gt;Nemotron 3 Ultra (high), Terminus-2&lt;/a&gt;: Geospatial Agent Index 68&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/codex-deepseek-v4-1-flash-max&quot;&gt;DeepSeek V4.1 Flash (max), Codex&lt;/a&gt;: Geospatial Agent Index 71&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/models/qwen-code-qwen-3-8-27b-xhigh&quot;&gt;Qwen 3.8 27B (xhigh), Qwen Code&lt;/a&gt;: Geospatial Agent Index 68&lt;/li&gt;&lt;/ul&gt;</content></entry><entry><title>Version 1.0 of the methodology</title><link href="https://www.axisspatial.com/arena/updates#2026-10-19-version-1-0-of-the-methodology"/><id>https://www.axisspatial.com/arena/updates#2026-10-19-version-1-0-of-the-methodology</id><updated>2026-10-19T00:00:00Z</updated><content type="html">&lt;ul&gt;&lt;li&gt;First public release.&lt;/li&gt;&lt;li&gt;The method is the same as in Pre-release 0.34: no task, weight or scoring rule changed.&lt;/li&gt;&lt;li&gt;&lt;a href=&quot;/arena/methodology#version-history&quot;&gt;Version history&lt;/a&gt;&lt;/li&gt;&lt;/ul&gt;</content></entry></feed>
