The story so far
Weights, every dated article, and repo commits on one rail.
17 repo commits since the weights landed
Teal is the weights (2026-09-20), grey ticks are repo commits.
this-that-model-1.0
1.88B parameters, text in, labels out, 256K tokens of context.
Verified 2026-10-25 Rank #0 in Decision models · score 0 method v1 Sources 3 Downloads 30d 830 · #598 of 619 tracked
Do people use it?830downloads, 30 days#598 of 619 we track · below the median+0 likes, downloads −51% this week18 likesHugging Face2026-10-24
Where it's used
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
No API provider serves it, so its downloads are people running it themselves: 1 quantized builds and 6 Spaces.
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 1.1 GB at INT4, 2.3 GB at FP8, 4.5 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
1 derivative on the Hub. Which build to run, by runtime: the most downloaded in each class.
1 derivative on the Hub
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
Benchmarks
15 rows
Benchmarks 15 rows
measured third-party leaderboard rowauthor-reported number from the model card or eval files (15)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported Calibration | 0.75 | — | model author | 2026-10-07 |
| author-reported Decomposition | 0.56 | 32×32 Whole state | model author | 2026-10-07 |
| author-reported Decomposition | 0.98 | 32×32 The part it depends on | model author | 2026-10-07 |
| author-reported Decomposition | 0.58 | 80×80 Whole state | model author | 2026-10-07 |
| author-reported Decomposition | 1 | 80×80 The part it depends on | model author | 2026-10-07 |
| author-reported Decomposition | 0.48 | 128×128 Whole state | model author | 2026-10-07 |
| author-reported Decomposition | 1 | 128×128 The part it depends on | model author | 2026-10-07 |
| author-reported Decomposition | 0.65 | 200×200 Whole state | model author | 2026-10-07 |
| author-reported Decomposition | 0.98 | 200×200 The part it depends on | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.84 | 2,250-question subset | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.87 | 2,250-question subset (seen) | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.83 | 2,250-question subset (unseen) | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.94 | the same cells as JSON (after) | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.94 | the same cells as prose (after) | model author | 2026-10-07 |
| author-reported limberc/this-that-spatial-bench | 0.78 | an ASCII block (after) | model author | 2026-10-07 |
What the card says
2 cited sentences
Summary generated, cited
It utilizes a Qwen3.5-style hybrid architecture featuring 18 DeltaNet linear attention layers and 6 full attention layers.[1][2] The model is released under an MIT license and generates zero output tokens during inference.[3][4]
- architectures
- “18 of 24 layers DeltaNet linear attention, 6 full attention”
- license
- “generating **zero** output tokens”
Generated 2026-10-07 15:19 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Use-case evidence
6 use cases
What it is good at 6 use cases
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
- Classify and make typed decisionsstructuraldecision-model estimate p=1.00tag: text-classificationpipeline: text-classification+1 more
- Run on a laptop or phonestructuraldecision-model estimate p=0.99fits: single-gpu
- Call tools and APIsstructuralchat template: <tool_call>
- Reason through hard problemsstructuralchat template: <think>
- Chat and follow instructionsstructuralchat template present
- Read long documentsstructuralconfig: 262,144 context
Facts
21 fields
Listed in Decision models
Facts
- this-that-model-1.0 has 1.88B parameters. (safetensors metadata, observed 2026-10-24)
- Context length is 262,144 tokens (256K). (config.json max_position_embeddings, observed 2026-10-24)
- License: MIT (friction: none). Permissive. (license file, observed 2026-10-24)
- Created 2026-09-20, last modified 2026-09-21. (HF API, observed 2026-10-24)
| Pipeline tag | text-classification |
|---|---|
| Library | transformers |
| Architectures | Qwen3_5ForCausalLM |
| Model type | qwen3_5_text |
| Dtype | bfloat16 |
| Params by dtype | BF16: 1,881,825,088 |
| Size band | S |
| MoE | no |
| Context length | 262,144 (config.json max_position_embeddings) |
| Languages | en |
| License | MIT (mit) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · Qwen3_5ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Qwen3_5ForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Hardware fit
3 estimates
Hardware fit ESTIMATE
Weights-only memory for this-that-model-1.0: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 4.5 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 2.3 GB · Edge (<4 GB) |
| INT4 (0.5 bytes/param) | 1.1 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
Decision facts
catalog
Decision-model catalog curated
| How it was made | fine-tune of Mapika decider-2b |
|---|---|
| Lineage bucket | D - derivative of another team's decision model |
| Lineage evidence | HF base_model |
| Base model | decider-2b |
| Hosted price (input) | self-hosted/M |
| Source | https://huggingface.co/flock-io/this-that-model-1.0 |
All variants
1 variant
Lineage
quantized: 1
Top variants (1 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| mradermacher/this-that-model-1.0-GGUF | quantized | GGUF | — | — | mit | mradermacher | 0 | 900 |
Repository changelog
17 commits
Repository changelog 17 commits
| Date | Commit | SHA |
|---|---|---|
| 2026-09-21 | Remove the limitations section, matching the repository | 3d92719 |
| 2026-09-21 | Add the seen/unseen split for Laya, measured on the same partition | 555bae4 |
| 2026-09-21 | Add Laya measured on our spatial benchmark, with what makes the number readable | e3ccef1 |
| 2026-09-20 | Upload README.md with huggingface_hub | 5ce6be4 |
| 2026-09-20 | Upload README.md with huggingface_hub | c9b7ac0 |
| 2026-09-20 | Upload model.safetensors with huggingface_hub | 424c9fa |
| 2026-09-20 | Upload README.md with huggingface_hub | cfa1794 |
| 2026-09-20 | Upload accuracy-vs-cost.png with huggingface_hub | 52b91d2 |
| 2026-09-20 | Upload README.md with huggingface_hub | 67f12d9 |
| 2026-09-20 | Upload chat_template.jinja with huggingface_hub | f77e3e8 |
| 2026-09-20 | Upload tokenizer_config.json with huggingface_hub | 36dd448 |
| 2026-09-20 | Upload tokenizer.json with huggingface_hub | 6c0f2a7 |
| 2026-09-20 | Upload model.safetensors with huggingface_hub | 20aeefe |
| 2026-09-20 | Upload generation_config.json with huggingface_hub | 03a5e3c |
| 2026-09-20 | Upload config.json with huggingface_hub | 284bf0b |
| 2026-09-20 | Upload README.md with huggingface_hub | 25a1493 |
| 2026-09-20 | initial commit | 92fdf9f |
Similar models
3 models
Similar models
- decider-2b · 1.88B same size band (S), same task (text-classification), permissive license
- mojev · 854M same size band (S), same task (text-classification), permissive license
- decider-4b · 4.21B same task (text-classification), permissive license
Sources
3 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:44 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:19 UTC · Last verified 2026-10-25
Follow this-that-model-1.0
Watch this model
- /feeds/models/flock-io--this-that-model-1.0.xml(RSS 2.0) · every change to this-that-model-1.0
- /feeds/families/flock-io--this-that.xml(RSS 2.0) · every change across the this-that family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.