Instruction Following · Benchmark profile

IFBench

Instruction-following evaluation measuring constraint and format reliability.

Not WeightedCurrent / rollingActive86 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on IFBench

Leaderboard · 86 models

MiniMax M3 leads at 82.9%, followed by Nemotron 3 Ultra (81.4%) and Grok 4.3 (81.3%).

Display onlyInstruction FollowingCurrent

#1Closed

MiniMax M3

MiniMax

MiniMax-M3

82.9%

Overall #23 · Context 1M tokens

#2Open

Nemotron 3 Ultra

NVIDIA

nemotron-3-ultra-550b

81.4%

Overall #70 · Context 262K tokens

#3Closed

Grok 4.3

xAI

grok-4.3

81.3%

Overall #40 · Context 1M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

MiniMaxNVIDIAxAIAlibaba CloudXiaomiAmazonOpenAIGoogleZ.AIMoonshot AIMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
MiniMax M3	#1	MiniMax	82.9%
Nemotron 3 Ultra	#2	NVIDIA	81.4%
Grok 4.3	#3	xAI	81.3%
Grok 4.20	#4	xAI	81.2%
Qwen3.7-Max	#5	Alibaba Cloud	80.5%
MiMo-V2.5-Pro	#6	Xiaomi	79.9%
Qwen3.7-Plus	#7	Alibaba Cloud	79.1%
Nova 2 Pro	#8	Amazon	79.0%
Qwen3.5 397B	#9	Alibaba Cloud	78.8%
GPT-5.2 Codex	#10	OpenAI	77.6%
Gemini 3.1 Flash-Lite	#11	Google	77.2%
Gemini 3.1 Pro Preview	#12	Google	77.1%
Qwen3.6 Max	#13	Alibaba Cloud	76.6%
Gemini 3.5 Flash	#14	Google	76.3%
GLM-5.1	#15	Z.AI	76.3%
Kimi K2.6	#16	Moonshot AI	76.0%
GPT-5.4 nano	#17	OpenAI	75.9%
GPT-5.5	#18	OpenAI	75.9%
Muse Spark	#19	Meta	75.9%
Qwen3.6 Plus	#20	Alibaba Cloud	75.8%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
MiniMax M3	#1	MiniMax	82.9%
Nemotron 3 Ultra	#2	NVIDIA	81.4%
Grok 4.3	#3	xAI	81.3%
Grok 4.20	#4	xAI	81.2%
Qwen3.7-Max	#5	Alibaba Cloud	80.5%
MiMo-V2.5-Pro	#6	Xiaomi	79.9%
Qwen3.7-Plus	#7	Alibaba Cloud	79.1%
Nova 2 Pro	#8	Amazon	79.0%
Qwen3.5 397B	#9	Alibaba Cloud	78.8%
GPT-5.2 Codex	#10	OpenAI	77.6%
Gemini 3.1 Flash-Lite	#11	Google	77.2%
Gemini 3.1 Pro Preview	#12	Google	77.1%
Qwen3.6 Max	#13	Alibaba Cloud	76.6%
Gemini 3.5 Flash	#14	Google	76.3%
GLM-5.1	#15	Z.AI	76.3%
Kimi K2.6	#16	Moonshot AI	76.0%
GPT-5.4 nano	#17	OpenAI	75.9%
GPT-5.5	#18	OpenAI	75.9%
Muse Spark	#19	Meta	75.9%
Qwen3.6 Plus	#20	Alibaba Cloud	75.8%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
MiniMax	1	82.9%
NVIDIA	1	81.4%
xAI	2	81.3%
Xiaomi	1	79.9%
Amazon	1	79.0%
Alibaba Cloud	5	78.2%
Google	3	76.9%
OpenAI	3	76.5%
Z.AI	1	76.3%
Moonshot AI	1	76.0%
Meta	1	75.9%

Benchmark score table (86 models)

One best score per model · higher is better

IFBench leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	MiniMax M3 MiniMax-M3	MiniMax	—	82.9%
#2	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	81.4%
#3	Grok 4.3 grok-4.3	xAI	closed	81.3%
#4	Grok 4.20 grok-4.20	xAI	closed	81.2%
#5	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	80.5%
#6	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	79.9%
#7	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	79.1%
#8	Nova 2 Pro nova-2-pro-preview	Amazon	closed	79.0%
#9	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	78.8%
#10	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	77.6%
#11	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	77.2%
#12	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	77.1%
#13	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	76.6%
#14	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	76.3%
#15	GLM-5.1 glm-5-1	Z.AI	open	76.3%
#16	Kimi K2.6 kimi-k2.6	Moonshot AI	open	76.0%
#17	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	75.9%
#18	GPT-5.5 gpt-5.5	OpenAI	closed	75.9%
#19	Muse Spark muse-spark	Meta	closed	75.9%
#20	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	75.8%
#21	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	75.7%
#22	MiniMax M2.7 minimax-m2-7	MiniMax	open	75.7%
#23	Gemma 4 31B gemma-4-31b	Google	open	75.6%
#24	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	75.6%
#25	GPT-5.2 gpt-5.2	OpenAI	closed	75.4%
#26	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	75.4%
#27	GPT-5 Codex gpt-5-codex	OpenAI	closed	74.1%
#28	Command A+ command-a-plus	Cohere	open	73.9%
#29	GPT-5.4 gpt-5.4	OpenAI	closed	73.9%
#30	Gemma 4 12B gemma-4-12b	Google	open	73.5%
#31	GLM-5.2 glm-5-2	Z.AI	open	73.3%
#32	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	73.3%
#33	GLM-5-Turbo glm-5-turbo	Z.AI	closed	73.2%
#34	GPT-5 gpt-5	OpenAI	closed	73.1%
#35	GPT-5.1 gpt.5.1	OpenAI	closed	72.9%
#36	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	72.7%
#37	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	72.5%
#38	Gemma 4 26B gemma-4-26b-a4b	Google	open	72.4%
#39	MiniMax M2 minimax-m2	MiniMax	open	72.3%
#40	GLM-5 glm-5	Z.AI	open	72.3%
#41	MiniMax M2.5 minimax-m2.5	MiniMax	open	71.6%
#42	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	71.5%
#43	o3 o3	OpenAI	closed	71.4%
#44	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	71.2%
#45	GPT-5 mini gpt-5-mini	OpenAI	closed	71.2%
#46	Qwen3 Max qwen3-max	Alibaba Cloud	closed	70.7%
#47	Gemini 3 Pro gemini-3-pro	Google	closed	70.4%
#48	o1 o1	OpenAI	closed	70.3%
#49	Kimi K2.5 kimi-k2.5	Moonshot AI	—	70.2%
#50	MiniMax M2.1 minimax-m2.1	MiniMax	open	69.9%
#51	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	68.8%
#52	o4-mini o4-mini	OpenAI	closed	68.7%
#53	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	68.1%
#54	GLM-4.7 glm-4.7	Z.AI	open	67.9%
#55	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	67.6%
#56	MiMo-V2.5 mimo-v2.5	Xiaomi	open	67.1%
#57	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	64.2%
#58	Claude Fable 5 claude-fable-5	Anthropic	closed	63.5%
#59	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	63.1%
#60	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	62.2%
#61	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	61.1%
#62	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	60.7%
#63	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	58%
#64	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	57.3%
#65	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	57.0%
#66	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	56.9%
#67	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	55.4%
#68	Gemini 3 Flash gemini-3-flash	Google	closed	55.1%
#69	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	54.7%
#70	Claude Opus 4 claude-opus-4	Anthropic	closed	53.7%
#71	Grok 4 grok-4	xAI	closed	53.7%
#72	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	53.5%
#73	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	52.3%
#74	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	51.2%
#75	Grok 4 Fast grok-4-fast	xAI	closed	50.5%
#76	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	48.7%
#77	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	48.3%
#78	Mistral Small 4 mistral-small-2603	Mistral AI	open	48.2%
#79	Grok 3 mini grok-3-mini	xAI	closed	45.9%
#80	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	44.6%
#81	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	43.6%
#82	GLM-4.6 glm-4.6	Z.AI	open	43.4%
#83	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	41.7%
#84	Grok Code Fast 1 grok-code-fast-1	xAI	closed	41.4%
#85	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	41.2%
#86	Mistral Large 3 mistral-large-2512	Mistral AI	open	36.2%

The top of this snapshot is led by MiniMax M3 at 82.9%; third place is 1.6 points behind. The top-10 spread is 5.3 points.

Compare top models

MiniMax M3 vs Nemotron 3 Ultra Nemotron 3 Ultra vs Grok 4.3 Grok 4.3 vs Grok 4.20 Grok 4.20 vs Qwen3.7-Max

About IFBench

Definition and scoring

Organisation: IFBench
Category: Instruction Following
Version: Not evaluated
Direction: higher is better
Ranking use: Not Weighted
Contamination risk: Unknown

Instruction-following evaluation measuring constraint and format reliability. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.

Open benchmark source ↗

FAQ

Quick answers

What does IFBench measure?

Instruction-following evaluation measuring constraint and format reliability.

Which model scores highest on IFBench?

MiniMax M3 by MiniMax currently leads with 82.9%.

How many models are evaluated on IFBench?

86 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

No. This benchmark is display-only and does not enter the overall Lumina composite.

IFEval

Rolling · 37 results · Not Weighted

Arena-Hard-Auto

Rolling · 1 results · Not Weighted

WildBench

Rolling · Not evaluated · Not Weighted

Instruction Following · Benchmark profile

IFBench

Instruction-following evaluation measuring constraint and format reliability.

Not WeightedCurrent / rollingActive86 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on IFBench

Leaderboard · 86 models

MiniMax M3 leads at 82.9%, followed by Nemotron 3 Ultra (81.4%) and Grok 4.3 (81.3%).

Display onlyInstruction FollowingCurrent

#1Closed

MiniMax M3

MiniMax

MiniMax-M3

82.9%

Overall #23 · Context 1M tokens

#2Open

Nemotron 3 Ultra

NVIDIA

nemotron-3-ultra-550b

81.4%

Overall #70 · Context 262K tokens

#3Closed

Grok 4.3

xAI

grok-4.3

81.3%

Overall #40 · Context 1M tokens

Visual analysis

See the field, not just the table.

Three views of the same sourced leaderboard: model placement, the score curve, and descriptive provider averages. These visuals never alter ranking calculations.

ModelsColours

MiniMaxNVIDIAxAIAlibaba CloudXiaomiAmazonOpenAIGoogleZ.AIMoonshot AIMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
MiniMax M3	#1	MiniMax	82.9%
Nemotron 3 Ultra	#2	NVIDIA	81.4%
Grok 4.3	#3	xAI	81.3%
Grok 4.20	#4	xAI	81.2%
Qwen3.7-Max	#5	Alibaba Cloud	80.5%
MiMo-V2.5-Pro	#6	Xiaomi	79.9%
Qwen3.7-Plus	#7	Alibaba Cloud	79.1%
Nova 2 Pro	#8	Amazon	79.0%
Qwen3.5 397B	#9	Alibaba Cloud	78.8%
GPT-5.2 Codex	#10	OpenAI	77.6%
Gemini 3.1 Flash-Lite	#11	Google	77.2%
Gemini 3.1 Pro Preview	#12	Google	77.1%
Qwen3.6 Max	#13	Alibaba Cloud	76.6%
Gemini 3.5 Flash	#14	Google	76.3%
GLM-5.1	#15	Z.AI	76.3%
Kimi K2.6	#16	Moonshot AI	76.0%
GPT-5.4 nano	#17	OpenAI	75.9%
GPT-5.5	#18	OpenAI	75.9%
Muse Spark	#19	Meta	75.9%
Qwen3.6 Plus	#20	Alibaba Cloud	75.8%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
MiniMax M3	#1	MiniMax	82.9%
Nemotron 3 Ultra	#2	NVIDIA	81.4%
Grok 4.3	#3	xAI	81.3%
Grok 4.20	#4	xAI	81.2%
Qwen3.7-Max	#5	Alibaba Cloud	80.5%
MiMo-V2.5-Pro	#6	Xiaomi	79.9%
Qwen3.7-Plus	#7	Alibaba Cloud	79.1%
Nova 2 Pro	#8	Amazon	79.0%
Qwen3.5 397B	#9	Alibaba Cloud	78.8%
GPT-5.2 Codex	#10	OpenAI	77.6%
Gemini 3.1 Flash-Lite	#11	Google	77.2%
Gemini 3.1 Pro Preview	#12	Google	77.1%
Qwen3.6 Max	#13	Alibaba Cloud	76.6%
Gemini 3.5 Flash	#14	Google	76.3%
GLM-5.1	#15	Z.AI	76.3%
Kimi K2.6	#16	Moonshot AI	76.0%
GPT-5.4 nano	#17	OpenAI	75.9%
GPT-5.5	#18	OpenAI	75.9%
Muse Spark	#19	Meta	75.9%
Qwen3.6 Plus	#20	Alibaba Cloud	75.8%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
MiniMax	1	82.9%
NVIDIA	1	81.4%
xAI	2	81.3%
Xiaomi	1	79.9%
Amazon	1	79.0%
Alibaba Cloud	5	78.2%
Google	3	76.9%
OpenAI	3	76.5%
Z.AI	1	76.3%
Moonshot AI	1	76.0%
Meta	1	75.9%

Benchmark score table (86 models)

One best score per model · higher is better

IFBench leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	MiniMax M3 MiniMax-M3	MiniMax	—	82.9%
#2	Nemotron 3 Ultra nemotron-3-ultra-550b	NVIDIA	open	81.4%
#3	Grok 4.3 grok-4.3	xAI	closed	81.3%
#4	Grok 4.20 grok-4.20	xAI	closed	81.2%
#5	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	80.5%
#6	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	79.9%
#7	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	79.1%
#8	Nova 2 Pro nova-2-pro-preview	Amazon	closed	79.0%
#9	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	78.8%
#10	GPT-5.2 Codex gpt-5.2-codex	OpenAI	closed	77.6%
#11	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	77.2%
#12	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	77.1%
#13	Qwen3.6 Max qwen3.6-max	Alibaba Cloud	closed	76.6%
#14	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	76.3%
#15	GLM-5.1 glm-5-1	Z.AI	open	76.3%
#16	Kimi K2.6 kimi-k2.6	Moonshot AI	open	76.0%
#17	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	75.9%
#18	GPT-5.5 gpt-5.5	OpenAI	closed	75.9%
#19	Muse Spark muse-spark	Meta	closed	75.9%
#20	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	75.8%
#21	Qwen3.5 122B qwen3.5-122b-a10b	Alibaba Cloud	open	75.7%
#22	MiniMax M2.7 minimax-m2-7	MiniMax	open	75.7%
#23	Gemma 4 31B gemma-4-31b	Google	open	75.6%
#24	Qwen3.5 27B qwen3.5-27b	Alibaba Cloud	open	75.6%
#25	GPT-5.2 gpt-5.2	OpenAI	closed	75.4%
#26	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	75.4%
#27	GPT-5 Codex gpt-5-codex	OpenAI	closed	74.1%
#28	Command A+ command-a-plus	Cohere	open	73.9%
#29	GPT-5.4 gpt-5.4	OpenAI	closed	73.9%
#30	Gemma 4 12B gemma-4-12b	Google	open	73.5%
#31	GLM-5.2 glm-5-2	Z.AI	open	73.3%
#32	GPT-5.4 mini gpt-5.4-mini	OpenAI	closed	73.3%
#33	GLM-5-Turbo glm-5-turbo	Z.AI	closed	73.2%
#34	GPT-5 gpt-5	OpenAI	closed	73.1%
#35	GPT-5.1 gpt.5.1	OpenAI	closed	72.9%
#36	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	72.7%
#37	Qwen3.5 35B qwen3.5-35b-a3b	Alibaba Cloud	open	72.5%
#38	Gemma 4 26B gemma-4-26b-a4b	Google	open	72.4%
#39	MiniMax M2 minimax-m2	MiniMax	open	72.3%
#40	GLM-5 glm-5	Z.AI	open	72.3%
#41	MiniMax M2.5 minimax-m2.5	MiniMax	open	71.6%
#42	Nemotron 3 Super nemotron-3-super-120b	NVIDIA	open	71.5%
#43	o3 o3	OpenAI	closed	71.4%
#44	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	71.2%
#45	GPT-5 mini gpt-5-mini	OpenAI	closed	71.2%
#46	Qwen3 Max qwen3-max	Alibaba Cloud	closed	70.7%
#47	Gemini 3 Pro gemini-3-pro	Google	closed	70.4%
#48	o1 o1	OpenAI	closed	70.3%
#49	Kimi K2.5 kimi-k2.5	Moonshot AI	—	70.2%
#50	MiniMax M2.1 minimax-m2.1	MiniMax	open	69.9%
#51	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	68.8%
#52	o4-mini o4-mini	OpenAI	closed	68.7%
#53	Kimi K2 Thinking kimi-k2-thinking	Moonshot AI	open	68.1%
#54	GLM-4.7 glm-4.7	Z.AI	open	67.9%
#55	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	67.6%
#56	MiMo-V2.5 mimo-v2.5	Xiaomi	open	67.1%
#57	MiMo-V2-Flash mimo-v2-flash	Xiaomi	open	64.2%
#58	Claude Fable 5 claude-fable-5	Anthropic	closed	63.5%
#59	Kimi K2.7 Code kimi-k2.7-code	Moonshot AI	open	63.1%
#60	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	62.2%
#61	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	61.1%
#62	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	60.7%
#63	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	58%
#64	Claude Sonnet 4.5 claude-sonnet-4-5	Anthropic	closed	57.3%
#65	DeepSeek V3.1 Terminus deepseek-v3.1-terminus	DeepSeek	open	57.0%
#66	Ling 2.6 1T ling-2.6-1t	InclusionAI	open	56.9%
#67	Claude Opus 4.1 claude-opus-4-1	Anthropic	closed	55.4%
#68	Gemini 3 Flash gemini-3-flash	Google	closed	55.1%
#69	Claude Sonnet 4 claude-sonnet-4	Anthropic	closed	54.7%
#70	Claude Opus 4 claude-opus-4	Anthropic	closed	53.7%
#71	Grok 4 grok-4	xAI	closed	53.7%
#72	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	53.5%
#73	Gemini 2.5 Flash gemini-2.5-flash	Google	closed	52.3%
#74	Qwen3.5 Omni Plus qwen3.5-omni-plus	Alibaba Cloud	closed	51.2%
#75	Grok 4 Fast grok-4-fast	xAI	closed	50.5%
#76	Gemini 2.5 Pro gemini-2.5-pro	Google	closed	48.7%
#77	Claude Sonnet 3.7 claude-3-7-sonnet	Anthropic	closed	48.3%
#78	Mistral Small 4 mistral-small-2603	Mistral AI	open	48.2%
#79	Grok 3 mini grok-3-mini	xAI	closed	45.9%
#80	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	44.6%
#81	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	43.6%
#82	GLM-4.6 glm-4.6	Z.AI	open	43.4%
#83	Kimi K2 0905 kimi-k2-0905	Moonshot AI	open	41.7%
#84	Grok Code Fast 1 grok-code-fast-1	xAI	closed	41.4%
#85	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	41.2%
#86	Mistral Large 3 mistral-large-2512	Mistral AI	open	36.2%

The top of this snapshot is led by MiniMax M3 at 82.9%; third place is 1.6 points behind. The top-10 spread is 5.3 points.

Compare top models

MiniMax M3 vs Nemotron 3 Ultra Nemotron 3 Ultra vs Grok 4.3 Grok 4.3 vs Grok 4.20 Grok 4.20 vs Qwen3.7-Max

About IFBench

Definition and scoring

Organisation: IFBench
Category: Instruction Following
Version: Not evaluated
Direction: higher is better
Ranking use: Not Weighted
Contamination risk: Unknown

Open benchmark source ↗

FAQ

Quick answers

What does IFBench measure?

Instruction-following evaluation measuring constraint and format reliability.

Which model scores highest on IFBench?

MiniMax M3 by MiniMax currently leads with 82.9%.

How many models are evaluated on IFBench?

86 models in the LuminaBench cohort have a qualifying score on this benchmark.

Does this affect overall Lumina rank?

No. This benchmark is display-only and does not enter the overall Lumina composite.

IFEval

Rolling · 37 results · Not Weighted

Arena-Hard-Auto

Rolling · 1 results · Not Weighted

WildBench

Rolling · Not evaluated · Not Weighted

Leaderboard · 86 models

MiniMax M3

Nemotron 3 Ultra

Grok 4.3

See the field, not just the table.

Benchmark score table (86 models)

Compare top models

Definition and scoring

Quick answers

What does IFBench measure?

Which model scores highest on IFBench?

How many models are evaluated on IFBench?

Does this affect overall Lumina rank?

More Instruction Following benchmarks

Leaderboard · 86 models

MiniMax M3

Nemotron 3 Ultra

Grok 4.3

See the field, not just the table.

Benchmark score table (86 models)

Compare top models

Definition and scoring

Quick answers

What does IFBench measure?

Which model scores highest on IFBench?

How many models are evaluated on IFBench?

Does this affect overall Lumina rank?

More Instruction Following benchmarks