Benchmarks for what frontier AI hasn't solved
Digital Work Index
An economically weighted measure of model performance across eight challenging benchmarks covering software engineering, logical reasoning and writing, and professional environments.
23.62
Fable 5.1
23.61
Opus 5.5
23.03
GPT-6 Astra
21.75
Opus 5
20.52
Grok 4.6
19.3
Grok 4.7
16.92
Gemini Flash 3.8
16.76
Kimi K3
16.45
GLM 5.3
16.29
Muse Spark 1.3
11.81
Qwen 3.8 Max
11.07
DeepSeek V4 Pro
6.95
Nemotron 3 Ultra 550B
All Models
Anthropic
Sep 22, 2026
Opus 5.5
RANK BY BENCHMARK
Agentic Coding 2.0
40.4%
#2/13
SWE-bench CLI
16.6%
#1/13
WorkplaceAgents
14.4%
#6/11
SnorkelLegal
31.4%
#7/13
SnorkelUnderwrite 2.0
22.7%
#7/13
xAI
Sep 21, 2026
Grok 4.7
RANK BY BENCHMARK
Agentic Coding 2.0
31.9%
#7/13
SWE-bench CLI
6.1%
#6/13
WorkplaceAgents
15.2%
#3/11
SnorkelLegal
32.8%
#6/13
SnorkelUnderwrite 2.0
30.5%
#1/13
OpenAI
Sep 03, 2026
GPT-6 Astra
RANK BY BENCHMARK
Agentic Coding 2.0
47.6%
#1/13
SWE-bench CLI
12.4%
#4/13
WorkplaceAgents
14.4%
#5/11
SnorkelLegal
20.9%
#13/13
SnorkelUnderwrite 2.0
21.2%
#8/13
Google DeepMind
Sep 02, 2026
Gemini Flash 3.8
RANK BY BENCHMARK
Agentic Coding 2.0
32.6%
#6/13
SWE-bench CLI
6.3%
#5/13
WorkplaceAgents
11.2%
#9/11
SnorkelLegal
25.3%
#12/13
SnorkelUnderwrite 2.0
19.1%
#11/13
All benchmarks
All Benchmarks
Sort: Newest
Benchmark Type: All
Science & Research
Open Benchmarks Grants
Terminal-Bench-Science
Evaluates agents on scientific workflows derived from researchers’ own work.
By Resolution Rate
1
Fable 5.1
40.0%
2
Opus 5
30.0%
3
GPT-5.6 Sol
22.4%
View archived benchmarks
