Files
2026-07-13 12:28:55 +08:00

83 lines
1.7 KiB
TOML

name = "Step 3.7 Flash"
description = "Newer StepFun flash model for faster agents, coding, and multimodal prompts"
release_date = "2026-05-29"
last_updated = "2026-05-29"
attachment = true
reasoning = true
temperature = true
tool_call = true
knowledge = "2026-01-01"
open_weights = true
[limit]
context = 256_000
input = 256_000
output = 256_000
[modalities]
input = ["text", "image"]
output = ["text"]
[[weights]]
label = "Hugging Face"
url = "https://huggingface.co/stepfun-ai/Step-3.7-Flash"
[[benchmarks]]
name = "SWE-Bench Pro"
score = 56.3
metric = "resolve rate"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "SWE-Bench Verified"
score = 76.5
metric = "resolved"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "Terminal-Bench"
score = 59.6
metric = "success rate"
version = "2.1"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "Humanity's Last Exam"
score = 47.2
metric = "accuracy"
variant = "with tools"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "BrowseComp"
score = 75.8
metric = "accuracy"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "Toolathlon"
score = 49.5
metric = "success rate"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "GDPval"
score = 45.8
metric = "wins or ties"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"
[[benchmarks]]
name = "ClawEval"
score = 67.1
metric = "pass^3"
version = "1.1"
source = "https://static.stepfun.com/blog/step-3.7-flash/"
date = "2026-05-29"