InclusionAI

Ling 3.1 Flash

Models.dev
Intelligence
29.6
Rank 153 of 813 measured models
Output speed
68.6 t/s
Rank 309 of 476 measured models
Price/1M
$0.45
Rank 298 of 750 measured models
Context window
1M
Rank 120 of 1420 measured models
Cost per task
$1.01
Rank 155 of 255 measured models

Ling 3.1 Flash is a text AI model from InclusionAI, released on Oct 1, 2026, with a context window of 1M tokens. Its weights are openly published. Ling 3.1 Flash scores 29.6 on the Artificial Analysis Intelligence Index, ranking 153rd of 813 measured models. The closest models on this index are Gemini 3.1 Pro Preview (29.7), Ling 3.1 Flash 1019 (29.6), and Qwen3.7 Max (29.5). At $0.45 per million tokens (3:1 input/output blended price), it is cheaper than Gemini 3.1 Pro Preview ($4.50) and Qwen3.7 Max ($3.75), priced close to Ling 3.1 Flash 1019 ($0.45). With an output speed of 68.6 t/s, it is slower than Gemini 3.1 Pro Preview (108 t/s), Ling 3.1 Flash 1019 (85.2 t/s), and Qwen3.7 Max (149 t/s). Every benchmark side by side: Ling 3.1 Flash vs Ling 3.1 Flash 1019.

Artificial Analysis Indices

Intelligence29.6
Rank 153 of 813 measured models

Capability indexes

Performance in specific industries, measured by Artificial Analysis
Finance & Accounting41.9
Rank 98 of 283 measured models
Strategy & Ops45.1
Rank 101 of 283 measured models
Legal38.2
Rank 111 of 283 measured models
Healthcare & Medical39.2
Rank 89 of 210 measured models
Engineering42.1
Rank 79 of 281 measured models
Economics47.7
Rank 75 of 289 measured models

Standard Benchmarks

HLE39.4%
Rank 101 of 720 measured models
SciCode54.1%
Rank 83 of 287 measured models
AA-LCR83.0%
Rank 28 of 631 measured models
Terminal-Bench 4.033.3%
Rank 53 of 250 measured models

Models at a similar level

The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy