Skip to content
Model Menu
Models
Compare
Leaderboard
Benchmarks
⌘K
Top
Chat
Coding
Agents
Writing
Vision
Image
Video
Speech
Embeddings
Decision
Music
Filters
Best for Agents & Tool Use
All models
2314
Chat & Knowledge
563
Coding
383
Agents & Tool Use
543
Writing & Language
361
Vision & Documents
397
Image Generation & Editing
184
Video Generation
148
Speech & Voice
358
Embeddings & Retrieval
688
Decision models
118
Music & Sound
29
AutomationBench-AA
GDPval-AA 2.1
Toolathlon-Verified
MCP-Atlas Public
MCP-Atlas full 1000
τ²-bench
AutomationBench · Zapier run
All benchmarks
83
Agents & Tool Use benchmarks
83
×
All
83
Agents
41
Health & law
1
Documents
1
Benchmark
By
Leader
Models
Agents
28
AutomationBench-AA
Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
AA
Gemini 4 Argon
77.5%
135
AutomationBench
Zapier run
End-to-end completion of business workflows using real SaaS applications
Zapier
Gemini 4 Argon
51.3%
2
+1 versions & slices
AutomationBench
Reported by the lab with its own setup
Lab
GLM 5.3
62.2
7
GDPval-AA (work tasks)
2.1
Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work
AA
Claude Opus 5.5
1,866
177
Toolathlon-Verified
Toolathlon-Verified as reported in this launch evaluation; the post is the authority for the recorded task and grading scope
GLM 5.3 Flash
78.4%
6
MCP-Atlas Public
MCPAtlas Public: MCPAtlas Public as reported in this launch evaluation; the post is the authority for the recorded task and grading scope
Muse Spark
82.4%
16
MCP-Atlas
full 1000
Tool orchestration across MCP servers on the stated task split
Scale
Muse Spark
82.2%
16
τ³-Banking
AA
Agents retrieve banking knowledge and execute tools to resolve customer-support workflows
AA
Grok 4.6
50.7%
134
τ³-bench
banking · 1.0.1
τ-bench official · 3-banking-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Sierra
Claude Opus 5
48.7%
21
AA Analyst Agent
Data-analyst tasks: the model explores data with tools and reports findings
AA
Claude Fable 5.1
57.5%
43
ITBench SRE
IT incident tasks: diagnose and fix broken systems like a site-reliability engineer
AA
GPT-5.6 Sol
56.2%
37
APEX-Agents
Agents complete long-horizon professional-services tasks across realistic applications
AA
Qwen3.8 Max (0902)
42.4%
29
Terminal-Bench Hard
Agents solve difficult software, system-administration and data-processing tasks in a terminal
AA
GPT-5.6 Sol
65.9%
239
Terminal-Bench
2.1
Older, easier version of Terminal-Bench; Still widely reported
AA
Claude Fable 5.1
91.4%
156
+1 versions & slices
4.0
Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Claude Sonnet 5.5
63.6%
134
Terminal-Bench Science
0.1
Agents complete expert-authored research workflows across five scientific domains in a terminal
AA
GPT-6 Astra
63.3%
45
AA-Briefcase
1.1 · Elo
Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating
AA
Claude Sonnet 5.5
1,812
142
DeepSWE
1.1
Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition
GPT-6 Astra
74.1%
44
EnterpriseOps-Gym-AA
Agents complete stateful enterprise workflows in eight domains; final database state determines success
AA
Claude Fable 5
51.1%
43
FrontierCode
1.1
Agents implement changes in real repositories, graded against maintainer-defined requirements
AA
Claude Opus 5.5
57.1%
38
Furniture Assembly
Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step
Epoch
Claude Opus 5.5
83.3%
31
EBR Bench
Agents learn to play Earthborne Rangers across repeated playthroughs
Epoch
GPT-6 Astra
76.2%
24
SWE-bench Pro (v2 full)
Agents resolve repository issues under the network-locked v2 protocol; full and 51-task Hard results are separate
Scale
Claude Opus 5
99.4%
16
SWE-bench Pro (v2 hard)
Agents resolve repository issues under the network-locked v2 protocol; full and 51-task Hard results are separate
Scale
Claude Opus 5
98%
16
SWE-bench Pro (private)
Agents resolve long-horizon professional repository tasks; public and private splits are separate
Scale
Muse Spark 1.1
51.5%
9
SpireClimb
Agents play Slay the Spire under the named game and difficulty setting
Epoch
GPT-6 Astra
25.7%
7
SpireClimb StS1 A20 Heart
Agents play Slay the Spire under the named game and difficulty setting
Epoch
GPT-6.1 Sol
20%
7
SpireClimb StS2 A10
Agents play Slay the Spire under the named game and difficulty setting
Epoch
GPT-6 Astra
34.7%
7
Health & law
1
Harvey LAB-AA
1.1
Legal agents work from case documents; every rubric criterion must pass and material hallucinations receive no credit
AA
Muse Wonder Full
13.1%
23
Documents
1
AA-LCR (long context)
1.1
Questions that need reading and connecting facts across very long documents (around 100K tokens)
AA
Kimi K3
88.7%
314
Reported by labs
40
GDPVal-AA
v2.1
Reported by the lab with its own setup
Lab
Kimi K3.1 Pro
1,846
1
OSWorld-Verified
Reported by the lab with its own setup
Lab
Claude Opus 4.8
83.4
5
OSWorld V2.1
Reported by the lab with its own setup
Lab
Kimi K3.1 Pro
79.6
1
Toolathlon
Reported by the lab with its own setup
Lab
GPT-5.4
54.6
6
MCP Atlas
Reported by the lab with its own setup
Lab
Muse Spark 1.1
88.1
5
BrowseComp
Reported by the lab with its own setup
Lab
Gemini 3.1 Pro
85.9
7
BrowseComp-zh
Reported by the lab with its own setup
Lab
DeepSeek V3.1 Terminus
45
1
Tau2-Bench
Reported by the lab with its own setup
Lab
Claude Sonnet 4.5
84.7
4
SWE-bench Pro (edition unstated)
Reported by the lab with its own setup
Lab
Claude Opus 5.5
89.9
44
DeepSWE (edition unstated)
Reported by the lab with its own setup
Lab
Claude Fable 5
70
11
Finance Agent
v2
Reported by the lab with its own setup
Lab
Muse Spark 1.1
57.2
11
HLE without tools
Reported by the lab with its own setup
Lab
Gemini 3.1 Pro
44.4
10
HLE with tools
Reported by the lab with its own setup
Lab
DeepSeek V4.1 Flash 1017
63.9
10
Agents Last Exam
Reported by the lab with its own setup
Lab
Kimi K3.1 Pro
43.2
7
Finch FinWorkBench
Reported by the lab with its own setup
Lab
Kimi K3
77.3
7
CyberGym-E2E AA
Reported by the lab with its own setup
Lab
Mistral Large 4
82
7
MCPAtlas Public
Reported by the lab with its own setup
Lab
Claude Opus 4.6
73.8
6
Automation-Bench
Reported by the lab with its own setup
Lab
DeepSeek V4.1 Flash 1017
54.8
6
Harvey Legal Agent Benchmark
Reported by the lab with its own setup
Lab
Mistral Large 4
15.8
6
B3 Attack Resistance
Reported by the lab with its own setup
Lab
Mistral Large 4
93.3
6
Harmful cyber request refusal mean
Reported by the lab with its own setup
Lab
Mistral Large 4
95.3
6
AA Briefcase combined Elo
Reported by the lab with its own setup
Lab
MiMo v2.6 Pro
1,516
6
KORA aggregate
Reported by the lab with its own setup
Lab
Mistral Large 4
1.7
6
JobBench
Reported by the lab with its own setup
Lab
Muse Spark 1.1
54.7
5
Humanity’s Last Exam with tools
Reported by the lab with its own setup
Lab
Muse Spark 1.1
62.1
5
DeepSearchQA
Reported by the lab with its own setup
Lab
GPT-5.5
87.8
5
CyberGym
Reported by the lab with its own setup
Lab
DeepSeek V4.1 Flash 1017
88.1
5
Cybench
Reported by the lab with its own setup
Lab
Mistral Large 4
93
5
MCP-Universe
Reported by the lab with its own setup
Lab
GPT-5
47.9
4
MCP-Mark
Reported by the lab with its own setup
Lab
GPT-5
50.9
4
Tool-Decathlon
Reported by the lab with its own setup
Lab
Claude Sonnet 4.5
38.6
4
Internal consumer agentic eval
Reported by the lab with its own setup
Lab
ChatGPT GPT-6
76.3
2
Consumer agentic eval first_s seconds
Reported by the lab with its own setup
Lab
ChatGPT GPT-5.6
72.1
2
Consumer agentic eval p25_s seconds
Reported by the lab with its own setup
Lab
ChatGPT GPT-5.6
75.2
2
Consumer agentic eval p50_s seconds
Reported by the lab with its own setup
Lab
ChatGPT GPT-5.6
78.3
2
Consumer agentic eval p75_s seconds
Reported by the lab with its own setup
Lab
ChatGPT GPT-6
85.4
2
Consumer agentic eval p100_s seconds
Reported by the lab with its own setup
Lab
ChatGPT GPT-6
94.2
2
SimpleQA search agent
Reported by the lab with its own setup
Lab
DeepSeek V3.1 Terminus
96.8
1
OfficeQA Pro
v2
Reported by the lab with its own setup
Lab
Kimi K3.1 Pro
64.4
1
Frontier Finance
Reported by the lab with its own setup
Lab
Kimi K3.1 Pro
58.6
1
Legacy · saturated or no longer reported
13
τ²-bench
AA
Tests service agents that use tools while coordinating with a simulated user
AA
GLM 5.2
99.1%
242
+8 versions & slices
airline · 0.1.3
τ-bench official · 2-airline-0.1.3: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Claude Sonnet 4.5
70%
8
retail · 0.1.3
τ-bench official · 2-retail-0.1.3: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Claude Sonnet 4.5
86.2%
8
overall · 0.1.3
τ-bench official · 2-overall-0.1.3: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Claude Sonnet 4.5
84.7%
7
telecom · 0.1.3
τ-bench official · 2-telecom-0.1.3: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Qwen3 Max Thinking
98.2%
7
airline · 1.0.1
τ-bench official · 2-airline-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Sierra
Claude Opus 4.5
84%
6
overall · 1.0.1
τ-bench official · 2-overall-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Sierra
Qwen3.5 397B A17B
87.9%
6
retail · 1.0.1
τ-bench official · 2-retail-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Sierra
Qwen3.5 397B A17B
84.4%
6
telecom · 1.0.1
τ-bench official · 2-telecom-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Sierra
Qwen3.5 397B A17B
97.8%
6
BFCL
v4
Tests function calling, multi-turn interactions, search and memory in the v4 suite
Berkeley
Claude Opus 4.5
77.5%
39
+2 versions & slices
v4 · prompt mode
Tests function calling, multi-turn interactions, search and memory in the v4 suite
Berkeley
Gemini 2.5 Flash
50.9%
16
v4 · native function calls
Tests function calling, multi-turn interactions, search and memory in the v4 suite
Berkeley
o3 (2025-04-16)
48.6%
2
SWE-bench Pro (public)
Agents resolve long-horizon professional repository tasks; public and private splits are separate
Muse Spark 1.1
61.5%
19
No benchmark matches
Model
AutomationBench · Zapier run
Per task
Price
Speed
Supports
Context
Gemini 4 Argon
Google
51.3
high
$1.62
$2
/
$10
79
1M
GPT-6 Astra
OpenAI
41.4
max
$3.26
$10
/
$50
66
1.05M
Show 541 not measured on AutomationBench · Zapier run
