Llama 4 Maverick Instruct vs GPT-6 Astra vs Qwen3.8 Max (0902) vs Gemini 4 Argon

Head to head on 13 boards: GPT-6 Astra places best on 7, Qwen3.8 Max (0902) places best on 3, Gemini 4 Argon places best on 3 and Llama 4 Maverick Instruct places best on 0. All 4 ranked on 0 of 18 boards.

Two at a time on this screen. Tap one to swap it in.

Llama 4 Maverick InstructGPT-6 AstraQwen3.8 Max (0902)Gemini 4 Argon

Board	Read	Llama 4 Maverick Instruct	GPT-6 Astra	Qwen3.8 Max (0902)	Gemini 4 Argon
Chat2
Text ArenaArena rating1dyesterday	Nov 91dyesterday		#311,475	#201,483	#21,523best of the picked
MultiChallengeScore1dyesterday	Nov 91dyesterday				#278.7%
Reasoning6
GPQA DiamondAccuracy1dyesterday	Nov 91dyesterday		#295.8%	#2692.7%	#196.4%best of the picked
Epoch ECIECI1dyesterday	Nov 91dyesterday		#1169.1best of the picked	#27157.4	#6166.4
ARC-AGI-2Score1dyesterday	Nov 91dyesterday	#840.0%	#195.0%best of the picked		#591.3%
AA IntelligenceIndex1dyesterday	Nov 91dyesterday		#356.9best of the picked	#1533.7	#750.4
HLEAccuracy1dyesterday	Nov 91dyesterday	#365.7%	#154.8%best of the picked
LiveBenchGlobal average1dyesterday	Nov 91dyesterday		#482.2%best of the picked	#1678.5%
Coding4
WebDev ArenaArena rating1dyesterday	Nov 91dyesterday		#31,770best of the picked	#121,677	#101,683
SWE-Bench ProResolved1dyesterday	Nov 91dyesterday		#490.2%		#392.2%best of the picked
SWE-bench VerifiedResolved · stale1dyesterday	Nov 91dyesterday	#3621.0%
Aider PolyglotPass rate · stale1dyesterday	Nov 91dyesterday	#4315.6%
Agents and tools2
Search ArenaArena rating1dyesterday	Nov 91dyesterday		#31,245best of the picked	#61,228
MCP AtlasPass rate1dyesterday	Nov 91dyesterday				#190.7%
Vision2
Vision ArenaArena rating1dyesterday	Nov 91dyesterday		#171,285	#21,300best of the picked
Document ArenaArena rating1dyesterday	Nov 91dyesterday		#111,477
Speed1
AA SpeedOutput speed1dyesterday	Nov 91dyesterday		#1621 tok/s	#260 tok/sbest of the picked	#1028 tok/s
Price1
AA PriceBlended price1dyesterday	Nov 91dyesterday		#17$20.0	#8$3.00best of the picked	#13$8.00

Green marks the best placing among the picked where two or more are ranked; ties are marked for each. Ranks are each board's own order (lower is better), scores as the board published them. Dash: not on that board.