Mistral Large 4 High vs Muse Spark vs GPT-6 Astra vs Gemini 3.8 Flash

Head to head on 14 boards: GPT-6 Astra places best on 9, Gemini 3.8 Flash places best on 4, Muse Spark places best on 1 and Mistral Large 4 High places best on 0. All 4 ranked on 0 of 16 boards.

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

Mistral Large 4 HighMuse SparkGPT-6 AstraGemini 3.8 Flash

Board	Read	Mistral Large 4 High	Muse Spark	GPT-6 Astra	Gemini 3.8 Flash
Chat2
Text ArenaArena ratingtodaytoday	Oct 7todaytoday		#121,489	#261,477	#71,495best of the picked
MultiChallengeScoretodaytoday	Oct 7todaytoday		#175.5%
Reasoning6
HLEAccuracytodaytoday	Oct 7todaytoday		#640.6%	#154.8%best of the picked	#444.5%
GPQA DiamondAccuracytodaytoday	Oct 7todaytoday		#4189.8%	#195.8%best of the picked	#395.4%
Epoch ECIECItodaytoday	Oct 7todaytoday		#44152.0	#2166.4best of the picked	#20156.7
AA IntelligenceIndextodaytoday	Oct 7todaytoday	#2138.4		#452.7best of the picked	#1740.9
LiveBenchGlobal averagetodaytoday	Oct 7todaytoday	#4970.5%		#482.2%best of the picked	#3275.8%
ARC-AGI-2Scoretodaytoday	Oct 7todaytoday			#195.0%best of the picked	#889.2%
Coding2
WebDev ArenaArena ratingtodaytoday	Oct 7todaytoday			#21,788best of the picked	#251,583
SWE-Bench ProResolvedtodaytoday	Oct 7todaytoday			#390.2%best of the picked	#958.8%
Agents and tools1
MCP AtlasPass ratetodaytoday	Oct 7todaytoday		#982.2%
Vision2
Vision ArenaArena ratingtodaytoday	Oct 7todaytoday		#61,294best of the picked	#181,284	#91,290
Document ArenaArena ratingtodaytoday	Oct 7todaytoday		#241,444	#131,468best of the picked
Speed1
AA SpeedOutput speedtodaytoday	Oct 7todaytoday	#10116 tok/s		#1862 tok/s	#4216 tok/sbest of the picked
Price1
AA PriceBlended pricetodaytoday	Oct 7todaytoday	#14$2.06		#26$20.0	#11$1.50best of the picked
Usage1
OpenRouterWeekly tokenstodaytoday	Oct 7todaytoday			#191.4T tok	#132.2T tokbest of the picked

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.