Atlas

Models

← All models

Qwen3.7-Plus (source unspecified)

Alibaba

This benchmark entry preserves evaluations reported under Qwen3.7-Plus or an unversioned provider alias whose dated checkpoint was not established by the source. Alibaba documents the current alias as equivalent to qwen3.7-plus-2026-05-26, but that mapping alone does not pin historical evaluations. Exact release timing, context, and input modalities are left unspecified; reported reasoning controls remain on each run configuration.

Benchmark scores

BenchmarkScore
A-Fantasia61.7
A-Fantasia (Valid Attempts with Retries)61.7
A-Fantasia Backwards Spelling74.0
A-Fantasia Chess45.0
A-Fantasia Chess (Valid Attempts with Retries)45.0
A-Fantasia Cube (Valid Attempts with Retries)66.0
A-Fantasia Cube Rotation66.0
A-Fantasia Spelling (Valid Attempts with Retries)74.0
AA-Briefcase Elo913
AA-Briefcase v1.1 Elo915
AA-LCR65.0
AA-LCR v1.173.0
AA-Omniscience Index1.1
AgentVQA48.6
AgentVQA Egocentric Videos60.0
AgentVQA Games61.5
AgentVQA Robotics39.0
AgentVQA Spatial Planning50.0
AgentVQA Web Agents32.5
APEX-Agents-AA (avg@3)22.4
Arena Document (Style Controlled)1453
Arena Text — Business, Management & Financial Ops (No Style Control)1435
Arena Text — Chinese (No Style Control)1509
Arena Text — Creative Writing (No Style Control)1438
Arena Text — English (No Style Control)1459
Arena Text — Entertainment, Sports & Media (No Style Control)1422
Arena Text — Exclude Ties (No Style Control)1454
Arena Text — Exclude Ties (Style Controlled)1458
Arena Text — Expert (No Style Control)1467
Arena Text — French (No Style Control)1473
Loading Atlas data…