Reflection vs Anthropic vs DeepSeek vs OpenAI

Head to head on 21 boards: Anthropic places best on 12, OpenAI places best on 8, Reflection places best on 1 and DeepSeek places best on 1 (a tie counts for each). All 4 ranked on 2 of 26 boards.

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

RReflectionAnthropicDeepSeekOpenAI

Board	Read	RReflection	Anthropic	DeepSeek	OpenAI
Chat2
MultiChallengeScoretodaytoday	Oct 21todaytoday		#667.7%Claude Haiku 5.5	#2646.1%DeepSeek V3P1	#182.0%best of the pickedGPT-6.1 Sol
Text ArenaArena ratingtodaytoday	Oct 21todaytoday		#21,507best of the pickedClaude Opus 5.5	#341,475DeepSeek V4.1 Flash	#171,485GPT-5.6 Sol
Reasoning7
ARC-AGI-2Scoretodaytoday	Oct 21todaytoday	#4623.5%Beam 1.0	#393.3%Claude Opus 5.5	#2172.9%DeepSeek V4.1 Flash	#195.0%best of the pickedGPT-6 Astra
Epoch ECIECItodaytoday	Oct 21todaytoday		#1168.8best of the pickedClaude Opus 5.5	#32155.9DeepSeek V4.1 Flash	#2168.8GPT-6 Astra
LiveBenchGlobal averagetodaytoday	Oct 21todaytoday		#183.4%best of the pickedClaude Fable 5.1	#881.1%DeepSeek V4.1 Flash	#482.2%GPT-6 Astra
GPQA DiamondAccuracytodaytoday	Oct 21todaytoday		#395.6%Claude Sonnet 5.5	#2591.7%DeepSeek V4 Pro 0813	#295.8%best of the pickedGPT-6 Astra
MMLU-ProAccuracy · staletodaytoday	Oct 21todaytoday		#489.1%Claude Opus 4.6	#3385.0%DeepSeek-V3.2	#389.3%best of the pickedo1
AA IntelligenceIndextodaytoday	Oct 21todaytoday		#158.0best of the pickedClaude Opus 5.5		#453.0GPT-6 Astra
HLEAccuracytodaytoday	Oct 21todaytoday		#246.5%Claude Fable 5.1		#154.8%best of the pickedGPT-6 Astra
Coding4
WebDev ArenaArena ratingtodaytoday	Oct 21todaytoday		#11,814best of the pickedClaude Opus 5.5	#191,619DeepSeek V4.1 Flash	#21,788GPT-6 Astra
SWE-bench VerifiedResolved · staletodaytoday	Oct 21todaytoday		#176.8%best of the pickedClaude Opus 4.5	#1170.0%DeepSeek-V3.2	#672.8%GPT-5.2
Aider PolyglotPass rate · staletodaytoday	Oct 21todaytoday		#1072.0%Claude Opus 4	#874.2%DeepSeek-V3.2	#188.0%best of the pickedGPT-5
SWE-Bench ProResolvedtodaytoday	Oct 21todaytoday		#198.0%best of the pickedClaude Opus 5		#490.2%GPT-6 Astra
Agents and tools3
BFCLOverall accuracy · staletodaytoday	Oct 21todaytoday		#177.5%best of the pickedClaude Opus 4.5	#1256.7%DeepSeek-V3.2	#763.0%o3
Search ArenaArena ratingtodaytoday	Oct 21todaytoday		#21,253Claude Opus 4.6		#11,257best of the pickedGPT-5.6 Sol
MCP AtlasPass ratetodaytoday	Oct 21todaytoday		#287.2%best of the pickedClaude Fable 5.1		#1181.8%GPT-5.6
Vision2
Vision ArenaArena ratingtodaytoday	Oct 21todaytoday		#11,308best of the pickedClaude Fable 5	#391,260DeepSeek V4.1 Flash	#61,294GPT-6.1 Sol
Document ArenaArena ratingtodaytoday	Oct 21todaytoday		#11,516best of the pickedClaude Opus 5		#61,486GPT-5.5
Image2
Text-to-ImageArena ratingtodaytoday	Oct 21todaytoday				#11,425GPT Image 2.5 Sunburst
Image EditArena ratingtodaytoday	Oct 21todaytoday				#11,524GPT Image 2.5 Sunburst
Video1
Text-to-VideoArena ratingtodaytoday	Oct 21todaytoday				#111,368Sora 2 Pro
Speech2
Open ASRAverage WERtodaytoday	Oct 21todaytoday				#306.38 WERgpt-transcribe (openai)
Speech ArenaArena ratingtodaytoday	Oct 21todaytoday				#391,103TTS 1 Hd
Speed1
AA SpeedOutput speedtodaytoday	Oct 21todaytoday	#6345 tok/sbest of the pickedBeam 1.0	#12236 tok/sClaude Haiku 5.5	#16209 tok/sDeepSeek V4.1 Flash	#9266 tok/sGPT OSS 20b
Price1
AA PriceBlended pricetodaytoday	Oct 21todaytoday		#3$0.20best of the pickedClaude Haiku 5.5	#8$0.53DeepSeek V4.1 Flash	#3$0.20best of the pickedGPT-6 Luna
Usage1
OpenRouterWeekly tokenstodaytoday	Oct 21todaytoday		#56.8T tokClaude Opus 5.5	#127.3T tokbest of the pickedDeepSeek V4.1 Flash	#66.2T tokGPT-6 Luna

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.