Agents · Benchmark profile

Claw-Eval

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks.

Reference2026Active33 models

Data verified 21 Jul 2026 · Methodology 1.6.0

Benchmark score on Claw-Eval

Leaderboard · 33 models

Ornith-1.0-397B leads at 77.1%, followed by MiniMax M3 (74.5%) and Qwen3.6 27B (72.4%).

Display onlyAgents2026

#1Open

Ornith-1.0-397B

DeepReinforce AI

ornith-1-0-397b

77.1%

Overall — · Context 256K tokens

#2Closed

MiniMax M3

MiniMax

MiniMax-M3

74.5%

Overall #16 · Context 1M tokens

#3Open

Qwen3.6 27B

Alibaba Cloud

qwen3.6-27b

72.4%

Overall #79 · Context 262K tokens

Visual analysis

See the field.

Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.

ViewModelsColours

DeepReinforce AIMiniMaxAlibaba CloudAnthropicStepFunXiaomiMetaZ.AIMoonshot AIOpenAIDeepSeek

Claw-Eval leaderboard

20 ranked models · higher is better

View accessible chart data

Accessible data for Claw-Eval leaderboard
Model	Rank	Provider	Score
Ornith-1.0-397B	#1	DeepReinforce AI	77.1%
MiniMax M3	#2	MiniMax	74.5%
Qwen3.6 27B	#3	Alibaba Cloud	72.4%
Claude Opus 4.6	#4	Anthropic	70.4%
Ornith-1.0-35B	#5	DeepReinforce AI	69.8%
Qwen3.6-35B-A3B	#6	Alibaba Cloud	68.7%
Claude Sonnet 4.6	#7	Anthropic	67.8%
Step 3.7 Flash	#8	StepFun	67.1%
Qwen3.7-Max	#9	Alibaba Cloud	65.2%
MiMo-V2.5-Pro	#10	Xiaomi	63.8%
Muse Spark	#11	Meta	63.8%
Ornith-1.0-9B	#12	DeepReinforce AI	63.1%
Qwen3.7-Plus	#13	Alibaba Cloud	62.7%
GLM-5.1	#14	Z.AI	62.3%
Kimi K2.6	#15	Moonshot AI	62.3%
MiMo-V2.5	#16	Xiaomi	62.3%
GPT-5.4	#17	OpenAI	60.3%
DeepSeek V4 Pro	#18	DeepSeek	59.8%
Claude Opus 4.5	#19	Anthropic	59.6%
Qwen3.6 Plus	#20	Alibaba Cloud	58.8%

Benchmark score table (33 models)

One best score per model · higher is better

Claw-Eval leaderboard ranked by score
Rank	Model	Provider	License	Evidence use	Score
#1	Ornith-1.0-397B ornith-1-0-397b	DeepReinforce AI	open	Estimated reference	77.1%
#2	MiniMax M3 MiniMax-M3	MiniMax	—	Reference only	74.5%
#3	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	Estimated reference	72.4%
#4	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	Reference only	70.4%
#5	Ornith-1.0-35B ornith-1-0-35b	DeepReinforce AI	open	Estimated reference	69.8%
#6	Qwen3.6-35B-A3B qwen3-6-35b-a3b	Alibaba Cloud	open	Estimated reference	68.7%
#7	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	Reference only	67.8%
#8	Step 3.7 Flash step-3-7-flash	StepFun	open	Estimated reference	67.1%
#9	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	Reference only	65.2%
#10	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	Reference only	63.8%
#11	Muse Spark muse-spark	Meta	closed	Reference only	63.8%
#12	Ornith-1.0-9B ornith-1-0-9b	DeepReinforce AI	open	Estimated reference	63.1%
#13	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	Reference only	62.7%
#14	GLM-5.1 glm-5-1	Z.AI	open	Reference only	62.3%
#15	Kimi K2.6 kimi-k2.6	Moonshot AI	open	Estimated reference	62.3%
#16	MiMo-V2.5 mimo-v2.5	Xiaomi	open	Estimated reference	62.3%
#17	GPT-5.4 gpt-5.4	OpenAI	closed	Reference only	60.3%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	Reference only	59.8%
#19	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	Reference only	59.6%
#20	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	Reference only	58.8%
#21	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	Estimated reference	57.8%
#22	Gemini 3.1 Pro gemini-3-1-pro	Google	closed	Estimated reference	57.8%
#23	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	Reference only	57.8%
#24	GLM-5 glm-5	Z.AI	open	Reference only	57.7%
#25	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	Estimated reference	56.8%
#26	GLM-5-Turbo glm-5-turbo	Z.AI	closed	Reference only	55.8%
#27	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	Reference only	53.8%
#28	Kimi K2.5 kimi-k2.5	Moonshot AI	—	Reference only	52.3%
#29	Gemini 3 Flash gemini-3-flash	Google	closed	Reference only	49.2%
#30	MiniMax M2.7 minimax-m2-7	MiniMax	open	Reference only	48.7%
#31	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	Reference only	45.2%
#32	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	Reference only	40.2%
#33	Nemotron 3 Super 100B nemotron-3-super-100b	NVIDIA	open	Estimated reference	5.5%

The top of this snapshot is led by Ornith-1.0-397B at 77.1%; third place is 4.7 points behind. The top-10 spread is 13.3 points.

Compare top models

Ornith-1.0-397B vs MiniMax M3 MiniMax M3 vs Qwen3.6 27B Qwen3.6 27B vs Claude Opus 4.6 Claude Opus 4.6 vs Ornith-1.0-35B

About Claw-Eval

Definition and scoring

Organisation: Bowen Ye, Rang Li, Qibin Yang, Yuanxin Liu, Linli Yao, Hanglong Lv, Zhihui Xie, Chenxin An, Lei Li, Lingpeng Kong, Qi Liu, Zhifang Sui, Tong Yang
Category: Agents
Version: 2026
Direction: higher is better
Ranking use: Reference
Contamination risk: Unknown

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks. 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 Claw-Eval measure?

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks.

Which model scores highest on Claw-Eval?

Ornith-1.0-397B by DeepReinforce AI currently leads with 77.1%.

How many models are evaluated on Claw-Eval?

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

Terminal-Bench

2.1 · 155 results · Weighted 1.6%

Berkeley Function-Calling Leaderboard

V3 · 2 results · Not Weighted

τ-bench

Rolling · 37 results · Weighted 1.6%

GAIA

Rolling · 13 results · Weighted 1.6%

BrowseComp

Rolling · 56 results · Weighted 1.6%

OSWorld

Rolling · 13 results · Weighted 1.6%

Agents · Benchmark profile

Claw-Eval

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks.

Reference2026Active33 models

Data verified 21 Jul 2026 · Methodology 1.6.0

Benchmark score on Claw-Eval

Leaderboard · 33 models

Ornith-1.0-397B leads at 77.1%, followed by MiniMax M3 (74.5%) and Qwen3.6 27B (72.4%).

Display onlyAgents2026

#1Open

Ornith-1.0-397B

DeepReinforce AI

ornith-1-0-397b

77.1%

Overall — · Context 256K tokens

#2Closed

MiniMax M3

MiniMax

MiniMax-M3

74.5%

Overall #16 · Context 1M tokens

#3Open

Qwen3.6 27B

Alibaba Cloud

qwen3.6-27b

72.4%

Overall #79 · Context 262K tokens

Visual analysis

See the field.

Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.

ViewModelsColours

DeepReinforce AIMiniMaxAlibaba CloudAnthropicStepFunXiaomiMetaZ.AIMoonshot AIOpenAIDeepSeek

Claw-Eval leaderboard

20 ranked models · higher is better

View accessible chart data

Accessible data for Claw-Eval leaderboard
Model	Rank	Provider	Score
Ornith-1.0-397B	#1	DeepReinforce AI	77.1%
MiniMax M3	#2	MiniMax	74.5%
Qwen3.6 27B	#3	Alibaba Cloud	72.4%
Claude Opus 4.6	#4	Anthropic	70.4%
Ornith-1.0-35B	#5	DeepReinforce AI	69.8%
Qwen3.6-35B-A3B	#6	Alibaba Cloud	68.7%
Claude Sonnet 4.6	#7	Anthropic	67.8%
Step 3.7 Flash	#8	StepFun	67.1%
Qwen3.7-Max	#9	Alibaba Cloud	65.2%
MiMo-V2.5-Pro	#10	Xiaomi	63.8%
Muse Spark	#11	Meta	63.8%
Ornith-1.0-9B	#12	DeepReinforce AI	63.1%
Qwen3.7-Plus	#13	Alibaba Cloud	62.7%
GLM-5.1	#14	Z.AI	62.3%
Kimi K2.6	#15	Moonshot AI	62.3%
MiMo-V2.5	#16	Xiaomi	62.3%
GPT-5.4	#17	OpenAI	60.3%
DeepSeek V4 Pro	#18	DeepSeek	59.8%
Claude Opus 4.5	#19	Anthropic	59.6%
Qwen3.6 Plus	#20	Alibaba Cloud	58.8%

Benchmark score table (33 models)

One best score per model · higher is better

Claw-Eval leaderboard ranked by score
Rank	Model	Provider	License	Evidence use	Score
#1	Ornith-1.0-397B ornith-1-0-397b	DeepReinforce AI	open	Estimated reference	77.1%
#2	MiniMax M3 MiniMax-M3	MiniMax	—	Reference only	74.5%
#3	Qwen3.6 27B qwen3.6-27b	Alibaba Cloud	open	Estimated reference	72.4%
#4	Claude Opus 4.6 claude-opus-4-6	Anthropic	closed	Reference only	70.4%
#5	Ornith-1.0-35B ornith-1-0-35b	DeepReinforce AI	open	Estimated reference	69.8%
#6	Qwen3.6-35B-A3B qwen3-6-35b-a3b	Alibaba Cloud	open	Estimated reference	68.7%
#7	Claude Sonnet 4.6 claude-sonnet-4-6	Anthropic	closed	Reference only	67.8%
#8	Step 3.7 Flash step-3-7-flash	StepFun	open	Estimated reference	67.1%
#9	Qwen3.7-Max qwen3.7-max	Alibaba Cloud	—	Reference only	65.2%
#10	MiMo-V2.5-Pro mimo-v2-5-pro	Xiaomi	closed	Reference only	63.8%
#11	Muse Spark muse-spark	Meta	closed	Reference only	63.8%
#12	Ornith-1.0-9B ornith-1-0-9b	DeepReinforce AI	open	Estimated reference	63.1%
#13	Qwen3.7-Plus qwen3.7-plus	Alibaba Cloud	—	Reference only	62.7%
#14	GLM-5.1 glm-5-1	Z.AI	open	Reference only	62.3%
#15	Kimi K2.6 kimi-k2.6	Moonshot AI	open	Estimated reference	62.3%
#16	MiMo-V2.5 mimo-v2.5	Xiaomi	open	Estimated reference	62.3%
#17	GPT-5.4 gpt-5.4	OpenAI	closed	Reference only	60.3%
#18	DeepSeek V4 Pro deepseek-v4-pro	DeepSeek	open	Reference only	59.8%
#19	Claude Opus 4.5 claude-opus-4-5	Anthropic	closed	Reference only	59.6%
#20	Qwen3.6 Plus qwen3-6-plus	Alibaba Cloud	closed	Reference only	58.8%
#21	DeepSeek V4 Flash deepseek-v4-flash	DeepSeek	open	Estimated reference	57.8%
#22	Gemini 3.1 Pro gemini-3-1-pro	Google	closed	Estimated reference	57.8%
#23	MiMo-V2-Pro mimo-v2-pro	Xiaomi	closed	Reference only	57.8%
#24	GLM-5 glm-5	Z.AI	open	Reference only	57.7%
#25	Qwen3.5 397B qwen3.5-397b-a17b	Alibaba Cloud	open	Estimated reference	56.8%
#26	GLM-5-Turbo glm-5-turbo	Z.AI	closed	Reference only	55.8%
#27	GLM-5V-Turbo glm-5v-turbo	Z.AI	closed	Reference only	53.8%
#28	Kimi K2.5 kimi-k2.5	Moonshot AI	—	Reference only	52.3%
#29	Gemini 3 Flash gemini-3-flash	Google	closed	Reference only	49.2%
#30	MiniMax M2.7 minimax-m2-7	MiniMax	open	Reference only	48.7%
#31	MiMo-V2-Omni mimo-v2-omni	Xiaomi	closed	Reference only	45.2%
#32	DeepSeek V3.2 deepseek-v3.2	DeepSeek	open	Reference only	40.2%
#33	Nemotron 3 Super 100B nemotron-3-super-100b	NVIDIA	open	Estimated reference	5.5%

The top of this snapshot is led by Ornith-1.0-397B at 77.1%; third place is 4.7 points behind. The top-10 spread is 13.3 points.

Compare top models

Ornith-1.0-397B vs MiniMax M3 MiniMax M3 vs Qwen3.6 27B Qwen3.6 27B vs Claude Opus 4.6 Claude Opus 4.6 vs Ornith-1.0-35B

About Claw-Eval

Definition and scoring

Organisation: Bowen Ye, Rang Li, Qibin Yang, Yuanxin Liu, Linli Yao, Hanglong Lv, Zhihui Xie, Chenxin An, Lei Li, Lingpeng Kong, Qi Liu, Zhifang Sui, Tong Yang
Category: Agents
Version: 2026
Direction: higher is better
Ranking use: Reference
Contamination risk: Unknown

Open benchmark source ↗

FAQ

Quick answers

What does Claw-Eval measure?

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks.

Which model scores highest on Claw-Eval?

Ornith-1.0-397B by DeepReinforce AI currently leads with 77.1%.

How many models are evaluated on Claw-Eval?

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

Terminal-Bench

2.1 · 155 results · Weighted 1.6%

Berkeley Function-Calling Leaderboard

V3 · 2 results · Not Weighted

τ-bench

Rolling · 37 results · Weighted 1.6%

GAIA

Rolling · 13 results · Weighted 1.6%

BrowseComp

Rolling · 56 results · Weighted 1.6%

OSWorld

Rolling · 13 results · Weighted 1.6%

Search LUMINA BENCH

Navigation

Claw-Eval

Leaderboard · 33 models

Ornith-1.0-397B

MiniMax M3

Qwen3.6 27B

See the field.

Claw-Eval leaderboard

Crop Claw-Eval leaderboard

Benchmark score table (33 models)

Crop Claw-Eval benchmark table

Compare top models

Definition and scoring

Quick answers

What does Claw-Eval measure?

Which model scores highest on Claw-Eval?

How many models are evaluated on Claw-Eval?

Claw-Eval

Leaderboard · 33 models

Ornith-1.0-397B

MiniMax M3

Qwen3.6 27B

See the field.

Claw-Eval leaderboard

Crop Claw-Eval leaderboard

Benchmark score table (33 models)

Crop Claw-Eval benchmark table

Compare top models

Definition and scoring

Quick answers

What does Claw-Eval measure?

Which model scores highest on Claw-Eval?

How many models are evaluated on Claw-Eval?

Leaderboard · 33 models

Ornith-1.0-397B

MiniMax M3

Qwen3.6 27B

See the field.

Benchmark score table (33 models)

Compare top models

Definition and scoring

Quick answers

What does Claw-Eval measure?

Which model scores highest on Claw-Eval?

How many models are evaluated on Claw-Eval?

More Agents benchmarks

Leaderboard · 33 models

Ornith-1.0-397B

MiniMax M3

Qwen3.6 27B

See the field.

Benchmark score table (33 models)

Compare top models

Definition and scoring

Quick answers

What does Claw-Eval measure?

Which model scores highest on Claw-Eval?

How many models are evaluated on Claw-Eval?

More Agents benchmarks