Text Leaderboard

Overall rankings for text generation models, powered by community votes and the Elo rating system.

Scroll horizontally to see all columns

Text

#	Model	Elo	Votes	Benchmark	$/1M In / Out	Value
1	MiniMax M3	1268	52	-	$0.30 / $1.20	1,691
2	GPT-5.2	1267	172	-	$1.75 / $14.00	161
3	Claude Haiku 4.5	1241	184	-	$1.00 / $5.00	414
4	GPT-5.6 Luna	1236	41	-	$0.20 / $1.20	1,766
5	Mistral Medium 3.5	1219	31	-	$1.50 / $7.50	271
6	Functionary Swahili Large	1216	157	67.7	$0.20 / $1.50	1,431
7	GLM 5.2	1208	48	-	$0.65 / $2.04	898
8	Claude Sonnet 4.5	1205	166	-	$3.00 / $15.00	134
9	GLM 5	1203	198	57.0	$0.60 / $1.92	955
10	Claude Sonnet 4.6	1201	174	-	$3.00 / $15.00	133
11	Functionary Swahili Mini	1191	77	62.2	$0.10 / $0.90	2,382
12	Qwen3.7 Max	1164	33	-	$1.48 / $4.42	395
13	GPT-5 Nano	1140	163	68.4	$0.05 / $0.40	5,067
14	DeepSeek V4 Flash	1118	44	-	$0.05 / $0.09	15,845
15	GPT-oss-120B	1117	162	59.1	$0.15 / $0.60	2,979
16Provisional rank	Kimi K2.6	1271	28	-	$0.95 / $4.00	514
17Provisional rankRank up 1 place	DeepSeek V4 Pro	1209	25	-	$0.42 / $0.84	1,924
18Provisional rankRank down 1 place	Step 3.7 Flash	1199	17	-	$0.20 / $1.15	1,776

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

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