wehub-resource-sync
98 lines
2.1 KiB
TOML
98 lines
2.1 KiB
TOML
name = "Gemini 3.5 Flash"
|
|
description = "Fast Gemini model balancing multimodal reasoning, tool use, and cost"
|
|
family = "gemini-flash"
|
|
release_date = "2026-05-19"
|
|
last_updated = "2026-05-19"
|
|
attachment = true
|
|
reasoning = true
|
|
temperature = true
|
|
tool_call = true
|
|
structured_output = true
|
|
knowledge = "2025-01"
|
|
open_weights = false
|
|
|
|
[limit]
|
|
context = 1_048_576
|
|
output = 65_536
|
|
|
|
[modalities]
|
|
input = ["text", "image", "video", "audio", "pdf"]
|
|
output = ["text"]
|
|
|
|
[[benchmarks]]
|
|
name = "Terminal-Bench"
|
|
score = 76.2
|
|
metric = "success rate"
|
|
harness = "Terminus-2"
|
|
version = "2.1"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "SWE-Bench Pro"
|
|
score = 55.1
|
|
metric = "resolve rate"
|
|
variant = "single attempt"
|
|
dataset = "public"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "MCP Atlas"
|
|
score = 83.6
|
|
metric = "success rate"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "Toolathlon"
|
|
score = 56.5
|
|
metric = "success rate"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "OSWorld-Verified"
|
|
score = 78.4
|
|
metric = "success rate"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "MMMU Pro"
|
|
score = 83.6
|
|
metric = "accuracy"
|
|
variant = "no tools"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "CharXiv Reasoning"
|
|
score = 84.2
|
|
metric = "accuracy"
|
|
variant = "no tools"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "Humanity's Last Exam"
|
|
score = 40.2
|
|
metric = "accuracy"
|
|
dataset = "full set, text + MM"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "ARC-AGI-2"
|
|
score = 72.1
|
|
metric = "accuracy"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|
|
|
|
[[benchmarks]]
|
|
name = "GDPval-AA"
|
|
score = 1656
|
|
metric = "Elo"
|
|
source = "https://deepmind.google/models/gemini/flash/"
|
|
date = "2026-05-19"
|