Z.ai vs Microsoft vs DeepSeek vs Anthropic

Head to head on 21 boards: Anthropic places best on 16, DeepSeek places best on 3, Z.ai places best on 1 and Microsoft places best on 1. All 4 ranked on 2 of 26 boards.

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

Z.aiMicrosoftDeepSeekAnthropic

Board	Read	Z.ai	Microsoft	DeepSeek	Anthropic
Chat2
Text ArenaArena rating1dyesterday	Nov 91dyesterday	#231,482GLM-5.3		#321,475DeepSeek V4.1 Flash	#41,505best of the pickedClaude Opus 5.5
MultiChallengeScore1dyesterday	Nov 91dyesterday			#574.9%best of the pickedDeepSeek V4P1 Flash	#1067.7%Claude Haiku 5.5
Reasoning7
MMLU-ProAccuracy · stale1dyesterday	Nov 91dyesterday	#2986.0%GLM-5	#7576.0%Phi 4 Plus	#3385.0%DeepSeek-V3.2	#489.1%best of the pickedClaude Opus 4.6
LiveBenchGlobal average1dyesterday	Nov 91dyesterday	#3176.1%GLM-5.3		#881.1%DeepSeek V4.1 Flash	#183.4%best of the pickedClaude Fable 5.1
AA IntelligenceIndex1dyesterday	Nov 91dyesterday	#1731.7GLM-5.3		#1632.6DeepSeek V4.1 Flash	#259.1best of the pickedClaude Opus 5.5
GPQA DiamondAccuracy1dyesterday	Nov 91dyesterday	#2891.9%GLM-5.2		#2193.2%DeepSeek V4.1 Flash	#395.6%best of the pickedClaude Sonnet 5.5
Epoch ECIECI1dyesterday	Nov 91dyesterday	#29156.9GLM-5.3		#19159.3DeepSeek V4.1 Flash	#3168.4best of the pickedClaude Opus 5.5
ARC-AGI-2Score1dyesterday	Nov 91dyesterday	#2865.8%GLM 5.3 Flash		#1978.1%DeepSeek V4.1 Flash	#393.3%best of the pickedClaude Opus 5.5
HLEAccuracy1dyesterday	Nov 91dyesterday	#288.3%GLM-4p5			#246.5%best of the pickedClaude Fable 5.1
Coding4
WebDev ArenaArena rating1dyesterday	Nov 91dyesterday	#251,621GLM-5.3		#141,665DeepSeek V4.1 Flash	#11,814best of the pickedClaude Opus 5.5
SWE-bench VerifiedResolved · stale1dyesterday	Nov 91dyesterday	#672.8%GLM-5		#1170.0%DeepSeek-V3.2	#176.8%best of the pickedClaude Opus 4.5
SWE-Bench ProResolved1dyesterday	Nov 91dyesterday	#1184.3%GLM-5.3			#198.0%best of the pickedClaude Opus 5
Aider PolyglotPass rate · stale1dyesterday	Nov 91dyesterday			#874.2%best of the pickedDeepSeek-V3.2	#1072.0%Claude Opus 4
Agents and tools3
BFCLOverall accuracy · stale1dyesterday	Nov 91dyesterday	#472.4%GLM 4.6	#5028.8%Phi 4	#1256.7%DeepSeek-V3.2	#177.5%best of the pickedClaude Opus 4.5
MCP AtlasPass rate1dyesterday	Nov 91dyesterday	#884.2%GLM-5.3			#387.2%best of the pickedClaude Fable 5.1
Search ArenaArena rating1dyesterday	Nov 91dyesterday				#51,229Claude Fable 5
Vision2
Vision ArenaArena rating1dyesterday	Nov 91dyesterday	#291,272GLM 5.3 Flash		#341,263DeepSeek V4.1 Flash	#11,309best of the pickedClaude Fable 5
Document ArenaArena rating1dyesterday	Nov 91dyesterday	#401,407GLM-5v Turbo			#11,499best of the pickedClaude Fable 5
Image2
Text-to-ImageArena rating1dyesterday	Nov 91dyesterday	#661,012GLM Image	#61,330best of the pickedMAI Image 2.6
Image EditArena rating1dyesterday	Nov 91dyesterday		#41,467MAI Image 2.7
Speech2
Open ASRAverage WER1dyesterday	Nov 91dyesterday		#35.55 WERazure-speech-09-2026 (microsoft)
Speech ArenaArena rating1dyesterday	Nov 91dyesterday		#311,145Azure Hd 2.5
Embeddings1
MTEBMean task score1dyesterday	Nov 91dyesterday		#174.3%harrier-oss-v1-27b (microsoft)
Speed1
AA SpeedOutput speed1dyesterday	Nov 91dyesterday	#539 tok/sbest of the pickedGLM 5.3 Flash		#1323 tok/sDeepSeek V4 Pro 0813	#931 tok/sClaude Sonnet 5.5
Price1
AA PriceBlended price1dyesterday	Nov 91dyesterday	#7$2.15GLM-5.3		#3$0.53DeepSeek V4.1 Flash	#1$0.20best of the pickedClaude Haiku 5.5
Usage1
OpenRouterWeekly tokens1dyesterday	Nov 91dyesterday	#213.2T tokGLM 5.3 Flash		#156.0T tokbest of the pickedDeepSeek V4.1 Flash	#66.8T tokClaude Opus 5.5

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.