Models
30 open decision models you can run locally with Ollaya, from their authors, pulled by name.
- winnowEldanRing
Decision models by EldanRing, fine-tuned from Google's Gemma 4 and published as GGUF
- 0.724
- accuracy
- 113 ms
- five questions
4.6b · 7.5b · 12b - layaConvai Innovations
Open decision models from Convai Innovations
- 0.357
- accuracy
- 18 ms
- five questions
322m · 421m - deciderMapika
Decoder decision models by Mapika on Qwen3.5: the answer is read from option-letter logits in one forward pass
- 0.678
- accuracy
- 197 ms
- five questions
0.75b · 1.9b · 2.2b · 4.2b - kevJared Palmer
Decision models by Jared Palmer: a LoRA on a Qwen3.5 base plus a pointer head that scores every option at its own span, in one forward pass per question
- 0.669
- accuracy
- 171 ms
- five questions
0.76b · 4.2b · 7.9b - nliMoritz Laurer
Zero-shot classifiers by Moritz Laurer: every option becomes a hypothesis scored for entailment
- 0.536
- accuracy
- 18 ms
- five questions
396m · 435m - gliclassKnowledgator
Instruction-following zero-shot classifiers by Knowledgator: all options of a question are scored in one pass, so on a GPU the cost barely grows with the number of options; on a CPU it grows with them
- 0.477
- accuracy
- 14 ms
- five questions
439m · 440m - qwen3guardQwen
Safety guard by the Qwen team: is a text safe, controversial or unsafe, and which unsafe category?
- —
- accuracy
- 30 ms
- five questions
0.6b - d1-omniLiquid AI
Liquid AI's d1-omni-600M, text questions: an LFM2.5 encoder with a decision head that scores every option at its own marker, all of a question's options in one pass, calibrated per question type and option count
- —
- accuracy
- —
- five questions
587m - decisionvLLM Semantic Router
Decision models by the vLLM Semantic Router contributors: a fully fine-tuned Qwen3.5 backbone plus an endpoint head that scores every option at its own last token against the question, in one forward pass per question
- 0.512
- accuracy
- 69 ms
- five questions
0.75b - vonVictor Hugo Panisa
Decision model by Victor Hugo Panisa on ModernBERT-large: every option is scored at its own marker, all options of a question in one pass, with an input-conditioned calibration
- 0.447
- accuracy
- 19 ms
- five questions
395m - jevk5alibiserikbay
Decision model by alibiserikbay, fine-tuned from Qwen3.5-4B and published as GGUF
- 0.629
- accuracy
- 95 ms
- five questions
4b - clmContrastive-LM
Contrastive decision model by Contrastive-LM: the Qwen3-8B encoder embeds the state and every option, and two trained heads pick the option closest to the state
- 0.357
- accuracy
- 95 ms
- five questions
8.2b - gliclass-v4Knowledgator
Knowledgator's GLiClass v4, the decoder-KV generation: a Qwen3.5 decoder (2B or 9B) reads the instructions, the state and the labels in one row, a two-layer scorer with up to three reasoning steps scores every label against the text, and the model's own calibrator rescales the logits
- —
- accuracy
- —
- five questions
2.4b · 9.5b - nimbleBespoke Labs
Decision model by Bespoke Labs: a LoRA on Qwen3.5-9B trained on contrastive pairs
- 0.665
- accuracy
- 1015 ms
- five questions
9b - jebfrontier-infra
Decision models by Jason Brashear and AINode: LoRAs merged into Qwen3.5 and Qwen3.8, published as GGUF
- —
- accuracy
- 111 ms
- five questions
4b · 9b · 27b - cygnetblockbrain-ai
Frozen Gemma 4 12B IT with blockbrain-ai's Cygnet prompt: the options as letters, one answer slot, and one calibration temperature
- 0.684
- accuracy
- 152 ms
- five questions
12b - jeevesPostHog
Decision model by PostHog: Qwen3.5-9B with a LoRA merged in and a pointer head that scores every option at its own marker
- —
- accuracy
- 388 ms
- five questions
9b - clefCloudflare
Decision models by Cloudflare
- 0.703
- accuracy
- 268 ms
- five questions
9b - snaplogitlab
logitlab's snap1-2b: MiniCPM5-2B fine-tuned to read one option letter after the prompt of emnlmn's snap engine
- 0.647
- accuracy
- 82 ms
- five questions
2b - decimaA. M. Madani
A
- 0.483
- accuracy
- 27 ms
- five questions
122m · 140m · 321m - arbiterCodekins · Zyot Lab
Codekins' Arbiter (Zyot Lab): Gemma 3 4B IT with a LoRA and a fixed 24-slot head that reads every option at the end of one prompt per question
- —
- accuracy
- —
- five questions
4.3b - credenceTxoka
Txoka's Credence v1: MiCA refinements of EldanRing's Winnow-E4B (Gemma 4), run from the author's Q8_0 GGUFs on llama.cpp
- —
- accuracy
- —
- five questions
7.5b - strands-deciderStrands Agents
Strands Agents' decision model: a LoRA and a pointer head on Qwen3.5-2B base or on Gemma 4 E2B / E4B
- —
- accuracy
- —
- five questions
2b · 4.6b · 8.9b - pplx-deciderPerplexity AI
Perplexity's decision model: Qwen3.5-4B distilled from its 27B decision models, with a readout over 255 answer codes
- —
- accuracy
- —
- five questions
4b - larghaPhocinae
Phocinae's Largha: Laya's decision network on mmBERT-small (144M), fine-tuned on the typed-decisions workflows
- —
- accuracy
- 18 ms
- five questions
144m - cobaOllaya
Ollaya's own decision model: Laya's network on Ettin encoders, distilled from pplx-decider:flash
- 0.591
- accuracy
- 21 ms
- five questions
1.1b · 75m · 164m · 421m - suraOllaya
Ollaya's own decoder decision model: Qwen3.5 distilled from pplx-decider:flash in its joint format
- 0.570
- accuracy
- 63 ms
- five questions
0.8b · 2b - d3vLLM Semantic Router
Decision 3.0 by the vLLM Semantic Router contributors: a fine-tuned Qwen3.5 whose full-attention layers read the whole prompt, with a readout over 255 answer codes
- 0.737
- accuracy
- 202 ms
- five questions
0.6b · 0.8b · 2b · 4b · 9b - decision2vLLM Semantic Router
Decision 2.0 by the vLLM Semantic Router contributors: a fully fine-tuned Qwen3 or Qwen3.5 backbone plus an endpoint head that scores every option at its own last token against the question, in one forward pass per question
- 0.680
- accuracy
- 251 ms
- five questions
0.60b · 0.75b · 1.88b · 4.21b · 7.94b - d1LiquidAI
LiquidAI d1-3B: typed decisions with max-pooled answer forms, Q8_0
- —
- accuracy
- —
- five questions
3b
Numbers are for each model's default tag, from the current Ollaya Bench edition. Typed-decisions accuracy is the argmax against the majority label on all 400 typed-decisions states; the labels have low annotator agreement, so compare models with each other rather than reading the numbers as absolutes. Latency is the median five-question request with a short state, end to end through the HTTP API on an RTX 5090 (qwen3guard: its four built-in questions). Larger models are more accurate and slower; see each model's page for how speed grows with the length of the state.
No models found.