Models/InclusionAI/Ling 3.0 Flash

Ling 3.0 Flash

InclusionAI Flash checkpoint. Open MIT weights. AA Intelligence Index 38. OpenRouter prints $0.021 / $0.063; AA prints $0.075 / $0.22. Those are two published tables. No official DeepSWE or vals card. Catalog max out is 32,768.

01

Identity

262,144Context in
32,768Context out
2026-07-23 (OpenRouter created 1784818580)Released
OpenWeights
Canonical name
Ling 3.0 Flash
Aliases
ling-3.0-flash, inclusionai/ling-3.0-flash
Continuum hosted id
Not listed as a Continuum hosted id
OpenRouter slug
inclusionai/ling-3.0-flash
Hugging Face
inclusionAI/Ling-3.0-flash
Modalities
text
Open weights Not on Continuum host list Vendor lab

Not listed as a Continuum hosted id. MIT on the official repo. OpenRouter catalog created unix 1784818580 (23 Jul 2026). 262,144 in / 32,768 out; text→text. Hugging Face id inclusionAI/Ling-3.0-flash.

HF config.json (fetched 19 Aug 2026): BailingMoeV3ForCausalLM, model_type bailing_hybrid, 42 layers, hidden 2560, 512 experts, 8 experts/token, 262,144 positions.

02

Should I use this for coding agents

Use Ling Flash when you want this official open checkpoint. AA Intelligence Index 38. We do not invent a DeepSWE percent. Cite both published dollar tables.

DeepSWE

No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.

omitted, not invented as of 2026-08-13 source
Artificial Analysispublished integer
38
bench
Artificial Analysis
version
Intelligence Index
harness
published integer
metric
Intelligence Index
value
38
independent as of 2026-08-19 source
Vals Index

We did not open a vals.ai card for this identity. Chip omitted.

omitted, not invented as of 2026-08-19 source

vals 404 as of 19 Aug 2026. Chip omitted.

Ranked view: Best coding models: the independent leaderboard puts this row and every other card that carries an independent coding score on one board, with the confidence intervals left visible.

Named benches, printed only

Copied from AA or vals HTML we opened. 0.0% placeholder bars are omitted. Do not average these into the hero tiles, and do not invent CI, $/task, or steps for Artificial Analysis.

BenchPrintedNoteSourceURLAs of
AA list pair $0.075 / $0.22 per 1M Printed AA In/Out. OpenRouter catalog is $0.021 / $0.063 plus $0.0042/M cache read. Artificial Analysis Ling 3.0 Flash source 2026-08-19
AA Index eval cost $72.81 Printed “it cost $72.81 to evaluate Ling 3.0 Flash on the Intelligence Index.” Artificial Analysis Ling 3.0 Flash source 2026-08-19
OpenRouter list pair $0.021 / $0.063 per 1M Catalog row. Not the AA pair. OpenRouter catalog source 2026-08-19
03

When not to use it

Trust this before you buy

  • No official DeepSWE or vals row (vals 404). 32k max out is tight for long agents. This is not a 128k–384k frontier card.
  • AA prints $0.075 / $0.22. OpenRouter prints $0.021 / $0.063 plus cache read $0.0042/M. Those are two published tables; do not collapse them, and do not invoice the cheaper pair as AA.
  • MIT open weights. No official GGUF on this repo. Community quants are community. Not a Continuum hosted id. 262k context is smaller than the 1M peers.

Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.

04

Artifact and Hugging Face downloads

Artifact

SPDX / license
MIT
Total params
HF safetensors.total: 127,486,405,600 stored tensors
Architecture
BailingMoeV3ForCausalLM · bailing_hybrid · 42 layers · 512 experts · 8 experts/tok · hidden 2560 · 262,144 positions
Native precision
HF tensors BF16 + F32
Files
24 safetensor shards (model-00001-of-00024 …)
Repo size
HF API usedStorage 254,993,303,374 bytes (237 GiB)
HF created
2026-08-02T16:14:41Z
Chat template
Jinja present on the Hub card.
Paper
No arXiv tag on the Hugging Face API payload we fetched.
Official HF repo
inclusionAI/Ling-3.0-flash
HF downloads
15,521
HF likes
351

Community quants: Community quants are community.

Official weights

This is the model repo, not Download for Mac. Downloads and likes from the Hugging Face API on 2026-08-19.

Hub file tree

Safetensor shards
24
Shard bytes
254,981,097,642 bytes
Other file bytes
18,404,594 bytes
Tree file count
39
Hub usedStorage
254,993,303,374 bytes

usedStorage and the sum of .safetensors sizes can differ (LFS pointers, non-shard files, duplicate copies). Cited separately. Source: recursive tree API, 2026-08-19.

PathBytes
model-00001-of-00024.safetensors10,715,426,912
model-00002-of-00024.safetensors10,522,791,950
model-00003-of-00024.safetensors10,522,792,184
model-00004-of-00024.safetensors10,711,541,790
model-00005-of-00024.safetensors10,522,791,950
model-00006-of-00024.safetensors10,522,793,948
model-00007-of-00024.safetensors10,522,794,860
model-00008-of-00024.safetensors10,711,544,514
model-00009-of-00024.safetensors10,522,794,626
model-00010-of-00024.safetensors10,522,794,626
model-00011-of-00024.safetensors10,522,794,860
model-00012-of-00024.safetensors10,711,544,514

Showing the first 12 of 24 safetensor shards. The tree also holds 15 non-shard files, none of them listed here. The snapshot was capped at 12 shards when the tree API was read on 2026-08-19, so the remaining 12 are on the Files tab, not on this page. Totals above are the full tree API sum, not this slice. Byte sizes are Hub tree API size fields, never invented.

05

Continuum serving

Not listed as a Continuum hosted id. OpenRouter inclusionai/ling-3.0-flash.

Continuum hosted id not on the host list

We list first-party and OpenRouter, plus Continuum only when the live public allowlist named the id. This is not a 15-host routing table.

06

Economics

OpenRouter catalog 19 Aug 2026: $0.021 / $0.063 per million, cache read $0.0042 /M, 262,144 / 32,768, created 1784818580. No DeepSWE $/task.

Artificial Analysis Ling 3.0 Flash page (19 Aug 2026) prints $0.075 / $0.22 and “it cost $72.81 to evaluate Ling 3.0 Flash on the Intelligence Index.” Cite both tables.

Price this model Opens the pricing calculator preloaded with Ling 3.0 Flash.

Artificial Analysis Intelligence Index 38, fetched 19 Aug 2026. Same page prints $0.075 / $0.22 and $72.81 to run the Index. The AA tile is the published integer 38. AA did not print a DeepSWE-style CI or step count we will copy onto the chip. Do not collapse AA’s pair into OpenRouter’s $0.021 / $0.063.

07

Provenance

ClaimSourceAs of
OR catalog $0.021 / $0.063, cache read $0.0042/M, 262,144 / 32,768, created 1784818580, hugging_face_id inclusionAI/Ling-3.0-flashOpenRouter catalog2026-08-19
HF safetensors.total 127,486,405,600; usedStorage 254,993,303,374; 24 shards; likes 351; downloads 15,521; created 2026-08-02T16:14:41Z; MITHugging Face API inclusionAI/Ling-3.0-flash2026-08-19
AA Index 38; printed $0.075 / $0.22; $72.81 to evaluate IndexArtificial Analysis Ling 3.0 Flash2026-08-19
vals 404vals.ai Ling 3.0 Flash2026-08-19
OpenRouter id inclusionai/ling-3.0-flash, context 262,144OpenRouter /api/v1/models2026-08-19
HF downloads 15,521Hugging Face API inclusionAI/Ling-3.0-flash2026-08-19
24 safetensor shards; shard bytes 254,981,097,642; tree files 39Hugging Face tree API inclusionAI/Ling-3.0-flash2026-08-19
Spaces API returned 11; official collection Ling 3.0 (4 items, 15 upvotes)Hugging Face Spaces / collections API2026-08-19
08

Compare, FAQ, and Get Plus

No other card in this cluster shares a published DeepSWE official row we can put next to this one. We will not compare on vendor-blog numbers.

Continue through InclusionAI's model family with Ling 2.6 1T.

FAQ

Why no DeepSWE chip?

The official DeepSWE board updated 13 Aug 2026 did not list Ling 3.0 Flash. Omitting is the rule.

Why two prices?

AA printed $0.075 / $0.22. OpenRouter printed $0.021 / $0.063 plus $0.0042/M cache. We cite both.

Why does the lab blog disagree with DeepSWE or Scale?

Lab posts pick a harness, an effort, a split, and sometimes a private eval set. DeepSWE publishes the official mini-swe-agent row with cost, tokens, and steps. Scale SWE-bench Pro only counts when the public shared-harness board has a row we can fetch. A higher lab number is usually a different test, not a better one.

OpenAI-compatible call

No Continuum hosted id is verified for this model, so this page does not invent a curl target. Use the first-party API or the OpenRouter slug inclusionai/ling-3.0-flash.

Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.