whatmodel.sh · find the best decision models you can run

Best Decision models for language understanding for a 64GB graphics card

A leaderboard of the best Decision models for language understanding for a 64GB graphics card (with 32GB of system RAM). The best one that runs on it right now is Clef-flash, followed by vLLM-SR d3-flash (9B) and vLLM-SR d3-mini (4B). Published typed-decision evaluations, with evaluator-specific protocols and tested settings retained. Each task score uses only its reviewed benchmarks. Models that run on it show the download that fits and the command to run it.

37 models · 21 benchmarks · updated

Compare tested decision models and inference recipes across text, vision and workflow tasks. Use task filters and raw results: probability consistency measures coherence, not correctness.

Filter by your hardware and budget → ~64GB fast · reset
ram[2][4][8][16][24][32][48][64][96][128][192][256]
[]
GB
gpu[apple silicon][graphics card][unified memory][none]
vram[4][8][12][16][24][32][48][64][80][96][128]
[]
GB
price[local][$0.05][$0.1][any]max per M input tokens
Filter by task
task[all tasks][tools & automation][retrieval & classification][language understanding][arts & human taste][knowledge & reasoning][vision][workflows][consistency]

Scores use reviewed language understanding evaluations only. Compare scores within this task. Models without task measurements are excluded.

New #1s for your hardware:
orRSS

detecting your OS…

$ terminal

detecting your OS…

Run curl "whatmodel.sh/decision?ram=32&gpu=card&vram=64&max=any"to get this list in your terminal (add &format=json for agents).

Score vs release date

Release dates available for 15 of 37 models. Undated models are omitted; month-only dates are plotted at the start of the month.

#modeltask score↓where↓RAM↓speed↓get itsources
#modeltask score↓where↓RAM↓speed↓get itsources
7Clef-omni Cloudflare73±2api $0.15/M input tokens~87GBdevelopers.cloudflare.com ↗Decision Index 0.4.1System One Bench (Ali Alhashimi)
14Clef Cloudflare71±1api $0.24/M input tokens +local~68GBdevelopers.cloudflare.com ↗Decision Index 0.4.1System One Bench (Ali Alhashimi)DecideBench (Cho Yin Yong)
17d1.1 (hosted) Liquid AI70±5api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
18Clef-flash · bf16 Cloudflare70±1local~25GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)DecideBench (Cho Yin Yong)
21d1-ultra Liquid AI69±5api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
31vLLM-SR d3-flash (9B) · ollaya-f32 vLLM Semantic Router67±1†local~18GBhow to run › Decision Index 0.4.1
32d1 (hosted) Liquid AI67±5api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
47vLLM-SR d3-mini (4B) · ollaya-f32 vLLM Semantic Router64±1†local~11GBhow to run › Decision Index 0.4.1
50Mercury Decide (free endpoint) Inception64±5api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
56Jebadiah 27B v2.1 (T1e) · bf16-t1e frontier-infra62±1†local (slow)~69GBhow to run › Decision Index 0.4.1
61Microsoft-Decision-1 Microsoft62±4api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
67Jev 1.13 TypeSafe61±1api $0.042/M input tokensdocs.typesafe.ai ↗Decision Index 0.4.1System One Bench (Ali Alhashimi)DecideBench (Cho Yin Yong)
68Span-02 (Gemma 4 26B-A4B) Respan61±1api price unknownopenrouter.ai ↗Decision Index 0.4.1System One Bench (Ali Alhashimi)
78Perplexity Decider Flash v1.2 (4B) · ollaya-f32 Perplexity60±1local~10GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
80Jev 1.14 TypeSafe60±4api $0.042/M input tokensdocs.typesafe.ai ↗System One Bench (Ali Alhashimi)
87Decider 4B · ollaya-f32 Mapika59±1†local~11GBhow to run › Decision Index 0.4.1
118GPT-6 Luna Decisions OpenAI56±3api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)DecideBench (Cho Yin Yong)
121vLLM-SR d3-nano (2B) · ollaya-f32 vLLM Semantic Router55±1local~8.2GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
129Bespoke Nimble 9B v3 · ollaya-f32 Bespoke Labs53±1†local~21GBhow to run › Decision Index 0.4.1
131Jebadiah 9B v2.1 (T1e) · bf16-t1e frontier-infra52±1†local~25GBhow to run › Decision Index 0.4.1
132vLLM-SR Decision 2.0 Lux 9B · ollaya-f32 vLLM Semantic Router52±1†local~20GBhow to run › Decision Index 0.4.1
141Bespoke Nimble 9B v2 · ollaya-f32 Bespoke Labs49±1†local~20GBhow to run › Decision Index 0.4.1
149baselabs joey-1 (2B) baselabs48±1api price unknownopenrouter.ai ↗Decision Index 0.4.1System One Bench (Ali Alhashimi)
151Solar Decide Upstage47±4api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
156GLiClass v4 edge (0.4B) · ollaya-f32 Knowledgator Engineering47±1†local~2.7GBhow to run › Decision Index 0.4.1
162vLLM-SR d3-lite (0.8B) · ollaya-f32 vLLM Semantic Router46±1local~4.5GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
164vLLM-SR Decision 2.0 Nox 4B · ollaya-f32 vLLM Semantic Router46±1†local~12GBhow to run › Decision Index 0.4.1
175Kev 4B r10 · ollaya-f32 Jared Palmer44±1local~11GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
197Solar Decide Flash Upstage41±4api price unknownopenrouter.ai ↗System One Bench (Ali Alhashimi)
207vLLM-SR Decision 2.0 Sol 2B · ollaya-f32 vLLM Semantic Router35±1†local~11GBhow to run › Decision Index 0.4.1
213LiquidAI d1-omni-600M · ollaya-f32 Liquid AI33±1local~2.7GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
214vLLM-SR d3-edge (0.6B) · ollaya-f32 vLLM Semantic Router33±1local~2.6GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
216Strands Decider 2B (Hobson v21) · bf16 Amazon Web Services32±1local~6.6GBhow to run › Decision Index 0.4.1
230vLLM-SR Decision 2.0 Kai 0.6B · ollaya-f32 vLLM Semantic Router26±1†local~3.7GBhow to run › Decision Index 0.4.1
231vLLM-SR Decision 2.0 Eos 0.8B · ollaya-f32 vLLM Semantic Router26±1†local~5.8GBhow to run › Decision Index 0.4.1
233Kev 0.8B r15 · ollaya-f32 Jared Palmer26±1local~4.6GBhow to run › Decision Index 0.4.1System One Bench (Ali Alhashimi)
236CLM-v0.1-8B Qwen3-8B heads · ollaya-f32 Contrastive-LM24±1local~15GBhow to run › Decision Index 0.4.1DecideBench (Cho Yin Yong)

+ 219 more don't fit your machine or budget · show them

256 of 256 catalogue models listed here · sources and licences

* new or thinly tested: ranked on too few benchmark results · † no independently run results verified · ~RAM is estimated; full context may need more memory

37 models · 21 benchmarks · updated