Leaderboard Overview

Rankings are powered by the Elo rating system — the same method used in competitive chess. Every time you vote in a head-to-head challenge, the winning model gains points and the losing model drops. The more challenges a model wins, the higher it climbs. Scores update in real time so you're always seeing the latest community consensus.

Text

#	Model(18/18)	Elo	Votes	Benchmark
1	MiniMax M3	1268	52	-
2	GPT-5.2	1267	172	-
3	Claude Haiku 4.5	1241	184	-
4	GPT-5.6 Luna	1236	41	-
5	Mistral Medium 3.5	1219	31	-
6Rank up 1 place	Functionary Swahili Large	1216	157	67.7
7Rank up 1 place	GLM 5.2	1208	48	-
8Rank down 2 places	Claude Sonnet 4.5	1205	166	-
9	GLM 5	1203	198	57.0
10	Claude Sonnet 4.6	1201	174	-

Arrows compare against an earlier weekly baseline, measured among models present in both weeks.

View all 18 models

Image

#	Model(10/10)	Elo	Votes
1	Nano Banana Pro (Gemini 3 Pro Image)	1358	384
2Rank up 1 place	Gemini 2.5 Flash Image (Nano Banana)	1344	989
3Rank down 1 place	Nano Banana 2 (Gemini 3.1 Flash Image)	1294	393
4Rank up 6 places	Riverflow V2.5 Fast	1233	324
5Rank up 1 place	FLUX.2 Klein 4B	1205	933
6Rank up 3 places	Riverflow V2 Fast	1170	1007
7Rank up 1 place	MAI-Image-2.5	1149	353
8Rank down 4 places	Seedream 4.5	1140	978
9Rank down 2 places	GPT Image 2	1122	372
10Rank down 5 places	FLUX.2 Max	1079	361

Arrows compare against an earlier weekly baseline, measured among models present in both weeks.

View all

TTS

#	Model(4/4)	Elo	Votes
1	MKA1 Speech TTS v1	1159	260
2	MKA1 Speech TTS v1.1	1146	262
3Provisional rank	Grok Voice TTS 1.0	1260	11
4Provisional rank	MAI-Voice-2	1235	9

Provisional — fewer than 30 votes. These models are ranked below every established model, so rank and Elo can disagree.

View all

STT

#	Model(8/8)	Elo	Votes
1	MKA1 Speech Transcription v1.1	1193	30
2Provisional rank	Parakeet TDT 0.6B V3	1229	13
3Provisional rank	Chirp 3	1221	14
4Provisional rank	Voxtral Mini Transcribe	1219	16
5Provisional rank	GPT-4o Transcribe	1189	11
6Provisional rank	GPT-4o Mini Transcribe	1189	18
7Provisional rank	MKA1 Speech Transcription v1	1185	24
8Provisional rank	Whisper Large V3 Turbo	1175	8

Provisional — fewer than 30 votes. These models are ranked below every established model, so rank and Elo can disagree.

View all