Text Arena: LLM Rankings & AI Model Benchmarks

Text Arena🏆Overall

View overall rankings across various AI models in text-to-text tasks across math, coding, creative writing, and other open-ended domains.

Oct 27, 2026

9,003,272 votes

419 models

	Rank Spread
113	93135	Mistral Large 4 Mistral · Proprietary	1432±6	12,066	$0.70 / $2.10	1M
122	102144	Mistral Medium 3.5 Mistral · Modified MIT	1426±6	11,791	$0.75 / $3.75	262.1K
146	129156	Mistral Large 3 Mistral · Apache 2.0	1414±3	88,325	$0.50 / $1.50	N/A
157	143166	Mistral Medium 3.1 (2508) Mistral · Proprietary	1408±3	95,269	$0.40 / $2	131.1K
186	170194	Mistral Medium 3 (2505) Mistral · Proprietary	1387±5	32,777	$0.40 / $2	131.1K
213	204234	Mistral Small 3.2 (2506) Mistral · Apache 2.0	1356±5	17,462	$0.10 / $0.30	262.1K
272	257286	Mistral Large 2 (2407) Mistral · Mistral Research	1314±4	45,459	$2 / $6	131.1K
284	271292	Mistral Large (2411) Mistral · MRL	1306±4	28,073	$2 / $6	128K
285	269293	Magistral Medium (2506) Mistral · Proprietary	1305±6	11,513	$2 / $5	40K
289	276293	Mistral Small 3.1 24B Instruct (2503) Mistral · Apache 2.0	1303±5	32,722	$0.10 / $0.30	32K
311	299317	Mistral Small 24B Instruct (2501) Mistral · Apache 2.0	1274±6	14,681	$0.05 / $0.08	32.8K
326	324334	Mistral Large (2402) Mistral · Proprietary	1243±5	62,436	$8 / $24	32K
329	324341	Ministral 8B (2410) Mistral · MRL	1238±9	4,781	$0.10 / $0.10	131.1K
335	328343	Mixtral 8x22B Instruct v0.1 Mistral · Apache 2.0	1230±5	51,416	$0.90 / $0.90	65.5K
342	333346	Mistral Medium Mistral · Proprietary	1223±5	34,550	$1.50 / $7.50	262.1K
352	347357	Mixtral 8x7B Instruct v0.1 Mistral · Apache 2.0	1197±4	73,503	$0.63 / $0.63	32K
366	354376	OpenHermes 2.5 Mistral 7B Apache-2.0	1176±10	5,006	$0.17 / $0.17	N/A
378	366395	Dolphin 2.2.1 Mistral 7B Apache-2.0	1152±15	1,679	$0.50 / $0.50	16.4K
381	373390	Mistral 7B Instruct v0.2 Mistral · Apache-2.0	1149±7	19,402	$0.20 / $0.20	32.8K
402	394404	Mistral 7B Instruct Mistral · Apache 2.0	1110±9	8,977	$0.07 / $0.28	4.1K

Text Arena🏆Overall

View overall rankings across various AI models in text-to-text tasks across math, coding, creative writing, and other open-ended domains.

Default Leaderboard Plots

Fraction of Model A Wins in A vs. B Battles (Ties Excluded)

Average Win Rate Against Each Opponent (Ties Excluded)

Battle Count for Each Pair of Models (Ties Excluded)

Confidence Intervals on Model Strength (via Bootstrapping)