Instruction Following · Benchmark profile

IFEval

An evaluation of whether language models follow verifiable natural-language instructions.

Not WeightedCurrent / rollingActive37 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on IFEval

Leaderboard · 37 models

Qwen3.7-Plus leads at 94.6%, followed by Qwen3.6 Plus (94.3%) and Qwen3.7-Max (94.3%).

Display onlyInstruction FollowingCurrent

#1Closed

Qwen3.7-Plus

Alibaba Cloud

qwen3.7-plus

94.6%

Overall #14 · Context 1M tokens

#2Closed

Qwen3.6 Plus

Alibaba Cloud

qwen3-6-plus

94.3%

Overall #37 · Context 1M tokens

#3Closed

Qwen3.7-Max

Alibaba Cloud

qwen3.7-max

94.3%

Overall #13 · 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

Alibaba CloudMoonshot AIZ.AIAnthropicMiniMaxxAIXiaomiGoogleOpenAIMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Qwen3.7-Plus	#1	Alibaba Cloud	94.6%
Qwen3.6 Plus	#2	Alibaba Cloud	94.3%
Qwen3.7-Max	#3	Alibaba Cloud	94.3%
Kimi K2.5	#4	Moonshot AI	93.9%
GLM-5	#5	Z.AI	92.6%
Claude Opus 4.5	#6	Anthropic	90.9%
MiniMax M3	#7	MiniMax	82.9%
Grok 4.3	#8	xAI	81.3%
MiMo-V2.5-Pro	#9	Xiaomi	79.9%
Gemini 3.1 Flash-Lite	#10	Google	77.2%
Gemini 3.1 Pro Preview	#11	Google	77.1%
Gemini 3.5 Flash	#12	Google	76.3%
GLM-5.1	#13	Z.AI	76.3%
GPT-5.4 nano	#14	OpenAI	75.9%
GPT-5.5	#15	OpenAI	75.9%
Muse Spark	#16	Meta	75.9%
MiniMax M2.7	#17	MiniMax	75.7%
Gemma 4 31B	#18	Google	75.6%
GPT-5.3-Codex	#19	OpenAI	75.4%
GPT-5.4	#20	OpenAI	73.9%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Qwen3.7-Plus	#1	Alibaba Cloud	94.6%
Qwen3.6 Plus	#2	Alibaba Cloud	94.3%
Qwen3.7-Max	#3	Alibaba Cloud	94.3%
Kimi K2.5	#4	Moonshot AI	93.9%
GLM-5	#5	Z.AI	92.6%
Claude Opus 4.5	#6	Anthropic	90.9%
MiniMax M3	#7	MiniMax	82.9%
Grok 4.3	#8	xAI	81.3%
MiMo-V2.5-Pro	#9	Xiaomi	79.9%
Gemini 3.1 Flash-Lite	#10	Google	77.2%
Gemini 3.1 Pro Preview	#11	Google	77.1%
Gemini 3.5 Flash	#12	Google	76.3%
GLM-5.1	#13	Z.AI	76.3%
GPT-5.4 nano	#14	OpenAI	75.9%
GPT-5.5	#15	OpenAI	75.9%
Muse Spark	#16	Meta	75.9%
MiniMax M2.7	#17	MiniMax	75.7%
Gemma 4 31B	#18	Google	75.6%
GPT-5.3-Codex	#19	OpenAI	75.4%
GPT-5.4	#20	OpenAI	73.9%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Alibaba Cloud	3	94.4%
Moonshot AI	1	93.9%
Anthropic	1	90.9%
Z.AI	2	84.4%
xAI	1	81.3%
Xiaomi	1	79.9%
MiniMax	2	79.3%
Google	4	76.6%
Meta	1	75.9%
OpenAI	4	75.3%

Benchmark score table (37 models)

One best score per model · higher is better

IFEval leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	94.6%
#2	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	94.3%
#3	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	94.3%
#4	Kimi K2.5 kimi-k2.5	Moonshot AI	—	93.9%
#5	GLM-5 glm-5	Z.AI	open	92.6%
#6	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	90.9%
#7	MiniMax M3 MiniMax-M3	MiniMax	—	82.9%
#8	Grok 4.3 grok-4.3	xAI	closed	81.3%
#9	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	79.9%
#10	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	77.2%
#11	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	77.1%
#12	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	76.3%
#13	GLM-5.1 glm-5-1	Z.AI	open	76.3%
#14	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	75.9%
#15	GPT-5.5 gpt-5.5	OpenAI	closed	75.9%
#16	Muse Spark muse-spark	Meta	closed	75.9%
#17	MiniMax M2.7 minimax-m2-7	MiniMax	open	75.7%
#18	Gemma 4 31B gemma-4-31b	Google	open	75.6%
#19	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	75.4%
#20	GPT-5.4 gpt-5.4	OpenAI	closed	73.9%
#21	GLM-5.2 glm-5-2	Z.AI	open	73.3%
#22	GLM-5-Turbo glm-5-turbo	Z.AI	closed	73.2%
#23	GPT-5.1 gpt.5.1	OpenAI	closed	72.9%
#24	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	72.7%
#25	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	71.2%
#26	Gemini 3 Pro gemini-3-pro	Google	closed	70.4%
#27	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	68.8%
#28	Claude Fable 5 claude-fable-5	Anthropic	closed	63.5%
#29	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	62.2%
#30	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	61.1%
#31	Gemini 3 Flash gemini-3-flash	Google	closed	55.1%
#32	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	53.5%
#33	Mistral Small 4 mistral-small-2603	Mistral AI	open	48.2%
#34	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	44.6%
#35	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	43.6%
#36	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	41.2%
#37	Mistral Large 3 mistral-large-2512	Mistral AI	open	36.2%

The top of this snapshot is led by Qwen3.7-Plus at 94.6%; third place is 0.30 points behind. The top-10 spread is 17.4 points.

Compare top models

Qwen3.7-Plus vs Qwen3.6 Plus Qwen3.6 Plus vs Qwen3.7-Max Qwen3.7-Max vs Kimi K2.5 Kimi K2.5 vs GLM-5

About IFEval

Definition and scoring

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

An evaluation of whether language models follow verifiable natural-language instructions. 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 IFEval measure?

An evaluation of whether language models follow verifiable natural-language instructions.

Which model scores highest on IFEval?

Qwen3.7-Plus by Alibaba Cloud currently leads with 94.6%.

How many models are evaluated on IFEval?

37 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.

Arena-Hard-Auto

Rolling · 1 results · Not Weighted

WildBench

Rolling · Not evaluated · Not Weighted

IFBench

Rolling · 86 results · Not Weighted

Instruction Following · Benchmark profile

IFEval

An evaluation of whether language models follow verifiable natural-language instructions.

Not WeightedCurrent / rollingActive37 models

Data verified 15 July 2026 · Methodology 1.2.0

Benchmark score on IFEval

Leaderboard · 37 models

Qwen3.7-Plus leads at 94.6%, followed by Qwen3.6 Plus (94.3%) and Qwen3.7-Max (94.3%).

Display onlyInstruction FollowingCurrent

#1Closed

Qwen3.7-Plus

Alibaba Cloud

qwen3.7-plus

94.6%

Overall #14 · Context 1M tokens

#2Closed

Qwen3.6 Plus

Alibaba Cloud

qwen3-6-plus

94.3%

Overall #37 · Context 1M tokens

#3Closed

Qwen3.7-Max

Alibaba Cloud

qwen3.7-max

94.3%

Overall #13 · 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

Alibaba CloudMoonshot AIZ.AIAnthropicMiniMaxxAIXiaomiGoogleOpenAIMeta

Leaderboard bars

20 ranked models · higher is better

View accessible chart data

Accessible data for Leaderboard bars
Model	Rank	Provider	Score
Qwen3.7-Plus	#1	Alibaba Cloud	94.6%
Qwen3.6 Plus	#2	Alibaba Cloud	94.3%
Qwen3.7-Max	#3	Alibaba Cloud	94.3%
Kimi K2.5	#4	Moonshot AI	93.9%
GLM-5	#5	Z.AI	92.6%
Claude Opus 4.5	#6	Anthropic	90.9%
MiniMax M3	#7	MiniMax	82.9%
Grok 4.3	#8	xAI	81.3%
MiMo-V2.5-Pro	#9	Xiaomi	79.9%
Gemini 3.1 Flash-Lite	#10	Google	77.2%
Gemini 3.1 Pro Preview	#11	Google	77.1%
Gemini 3.5 Flash	#12	Google	76.3%
GLM-5.1	#13	Z.AI	76.3%
GPT-5.4 nano	#14	OpenAI	75.9%
GPT-5.5	#15	OpenAI	75.9%
Muse Spark	#16	Meta	75.9%
MiniMax M2.7	#17	MiniMax	75.7%
Gemma 4 31B	#18	Google	75.6%
GPT-5.3-Codex	#19	OpenAI	75.4%
GPT-5.4	#20	OpenAI	73.9%

Score curve

Shows how quickly scores separate down the field

View accessible chart data

Accessible data for Score curve
Model	Rank	Provider	Score
Qwen3.7-Plus	#1	Alibaba Cloud	94.6%
Qwen3.6 Plus	#2	Alibaba Cloud	94.3%
Qwen3.7-Max	#3	Alibaba Cloud	94.3%
Kimi K2.5	#4	Moonshot AI	93.9%
GLM-5	#5	Z.AI	92.6%
Claude Opus 4.5	#6	Anthropic	90.9%
MiniMax M3	#7	MiniMax	82.9%
Grok 4.3	#8	xAI	81.3%
MiMo-V2.5-Pro	#9	Xiaomi	79.9%
Gemini 3.1 Flash-Lite	#10	Google	77.2%
Gemini 3.1 Pro Preview	#11	Google	77.1%
Gemini 3.5 Flash	#12	Google	76.3%
GLM-5.1	#13	Z.AI	76.3%
GPT-5.4 nano	#14	OpenAI	75.9%
GPT-5.5	#15	OpenAI	75.9%
Muse Spark	#16	Meta	75.9%
MiniMax M2.7	#17	MiniMax	75.7%
Gemma 4 31B	#18	Google	75.6%
GPT-5.3-Codex	#19	OpenAI	75.4%
GPT-5.4	#20	OpenAI	73.9%

Provider profile

Average within the visible cohort · descriptive only

View accessible chart data

Accessible data for Provider profile
Provider	Models	Average score
Alibaba Cloud	3	94.4%
Moonshot AI	1	93.9%
Anthropic	1	90.9%
Z.AI	2	84.4%
xAI	1	81.3%
Xiaomi	1	79.9%
MiniMax	2	79.3%
Google	4	76.6%
Meta	1	75.9%
OpenAI	4	75.3%

Benchmark score table (37 models)

One best score per model · higher is better

IFEval leaderboard ranked by score
Rank	Model	Provider	License	Score
#1	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	94.6%
#2	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	94.3%
#3	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	94.3%
#4	Kimi K2.5 kimi-k2.5	Moonshot AI	—	93.9%
#5	GLM-5 glm-5	Z.AI	open	92.6%
#6	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	90.9%
#7	MiniMax M3 MiniMax-M3	MiniMax	—	82.9%
#8	Grok 4.3 grok-4.3	xAI	closed	81.3%
#9	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	79.9%
#10	Gemini 3.1 Flash-Lite gemini-3.1-flash-lite	Google	closed	77.2%
#11	Gemini 3.1 Pro Preview gemini-3.1-pro-preview	Google	closed	77.1%
#12	Gemini 3.5 Flash gemini-3.5-flash	Google	closed	76.3%
#13	GLM-5.1 glm-5-1	Z.AI	open	76.3%
#14	GPT-5.4 nano gpt.5.4.nano	OpenAI	closed	75.9%
#15	GPT-5.5 gpt-5.5	OpenAI	closed	75.9%
#16	Muse Spark muse-spark	Meta	closed	75.9%
#17	MiniMax M2.7 minimax-m2-7	MiniMax	open	75.7%
#18	Gemma 4 31B gemma-4-31b	Google	open	75.6%
#19	GPT-5.3-Codex gpt-5.3-codex	OpenAI	closed	75.4%
#20	GPT-5.4 gpt-5.4	OpenAI	closed	73.9%
#21	GLM-5.2 glm-5-2	Z.AI	open	73.3%
#22	GLM-5-Turbo glm-5-turbo	Z.AI	closed	73.2%
#23	GPT-5.1 gpt.5.1	OpenAI	closed	72.9%
#24	GPT-5.6 Sol gpt-5.6-sol	OpenAI	closed	72.7%
#25	GPT-5.6 Terra gpt-5.6-terra	OpenAI	closed	71.2%
#26	Gemini 3 Pro gemini-3-pro	Google	closed	70.4%
#27	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	68.8%
#28	Claude Fable 5 claude-fable-5	Anthropic	closed	63.5%
#29	Claude Opus 4.8 claude-opus-4-8	Anthropic	closed	62.2%
#30	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	61.1%
#31	Gemini 3 Flash gemini-3-flash	Google	closed	55.1%
#32	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	53.5%
#33	Mistral Small 4 mistral-small-2603	Mistral AI	open	48.2%
#34	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	44.6%
#35	Claude Opus 4.7 claude-opus-4-7	Anthropic	closed	43.6%
#36	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	41.2%
#37	Mistral Large 3 mistral-large-2512	Mistral AI	open	36.2%

The top of this snapshot is led by Qwen3.7-Plus at 94.6%; third place is 0.30 points behind. The top-10 spread is 17.4 points.

Compare top models

Qwen3.7-Plus vs Qwen3.6 Plus Qwen3.6 Plus vs Qwen3.7-Max Qwen3.7-Max vs Kimi K2.5 Kimi K2.5 vs GLM-5

About IFEval

Definition and scoring

Organisation: Google Research
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 IFEval measure?

An evaluation of whether language models follow verifiable natural-language instructions.

Which model scores highest on IFEval?

Qwen3.7-Plus by Alibaba Cloud currently leads with 94.6%.

How many models are evaluated on IFEval?

37 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.

Arena-Hard-Auto

Rolling · 1 results · Not Weighted

WildBench

Rolling · Not evaluated · Not Weighted

IFBench

Rolling · 86 results · Not Weighted

Leaderboard · 37 models

Qwen3.7-Plus

Qwen3.6 Plus

Qwen3.7-Max

See the field, not just the table.

Benchmark score table (37 models)

Compare top models

Definition and scoring

Quick answers

What does IFEval measure?

Which model scores highest on IFEval?

How many models are evaluated on IFEval?

Does this affect overall Lumina rank?

More Instruction Following benchmarks

Leaderboard · 37 models

Qwen3.7-Plus

Qwen3.6 Plus

Qwen3.7-Max

See the field, not just the table.

Benchmark score table (37 models)

Compare top models

Definition and scoring

Quick answers

What does IFEval measure?

Which model scores highest on IFEval?

How many models are evaluated on IFEval?

Does this affect overall Lumina rank?

More Instruction Following benchmarks