Best Open-Weights Decision Models

Open-weights decision models can run next to your data, be fine-tuned and audited. Models are ranked by their Full score on the Jev Decision Index, a third-party board of 37 benchmarks with private held-out sets, until scores from our own Metatext Decision Benchmark are published.

Ranked list (top 25)

#	Model	Decision Index	ECE	p50 latency	Per 1k decisions	Params
1	Lychee-27B LycheeAIOpen weights	67.3	0.108	97 ms	—	26B
2	vLLM-SR d3 (27B) vLLM Semantic RouterOpen weights	67.2	0.043	97 ms	—	26B
3	linne-oak-1-0 MISA-AICOREOpen weights	66.9	0.105	105 ms	—	31B
4	Decider 31B (NVFP4) MapikaOpen weights	66.0	0.024	72 ms	—	33B
5	bobcat-1.3 sanghwa-naOpen weights	64.8	0.102	64 ms	—	31B
6	amos-1-large Scira (mukaddamzaid)Open weights	64.8	0.032	148 ms	—	—
7	frontier-infra Jebadiah 27B v2.1 Frontier InfraOpen weights	64.4	0.056	303 ms	—	28B
8	Sieve-31B Sthānika AIOpen weights	64.3	0.033	253 ms	—	31B
9	AnyJev L0 · Gemma-4-31B morriszjmOpen weights	63.6	0.240	98 ms	—	31B
10	deck31b (Gemma 4 31B) krishna-gogineni-765Open weights	63.0	0.204	219 ms	—	33B
11	Darwin-27B-ZTC FINAL_BenchOpen weights	63.0	0.045	104 ms	—	26B
12	KnowLine-35B-A3B-Gen1 PelaAIOpen weights	62.9	0.048	41 ms	—	36B
13	Perplexity Decider v1.1 (27B) PerplexityOpen weights	62.8	0.071	104 ms	$0.0040	28B
14	decisio 0.6.0 (Gemma 4 31B) aminryOpen weights	62.7	0.142	54 ms	—	31B
15	Decider chat · Gemma-4-31B MapikaOpen weights	61.6	0.047	109 ms	—	33B
16	Torchcast Decision 27B Torchcast AIOpen weights	61.5	0.024	100 ms	—	28B
17	Qwen Decision 27B noncausal LoRA jsaurabhOpen weights	61.1	0.051	127 ms	—	28B
18	Quyet-1.0-Large chinhncOpen weights	60.8	0.078	105 ms	—	33B
19	Sieve 27B Sthānika AIOpen weights	60.2	0.065	101 ms	—	28B
20	OneJev1.5-27B OmniJevOpen weights	59.9	0.024	224 ms	—	27B
21	mogan-decision-31B-it Mogan AIOpen weights	59.8	0.054	109 ms	—	31B
22	Shisa DE-2 (repeat) Shisa.AIOpen weights	59.8	0.207	57 ms	—	26B
23	linne-fern 1.0 (27B) MISA-AICOREOpen weights	59.6	0.079	108 ms	—	28B
24	Kev 27B Jared PalmerOpen weights	59.6	0.022	135 ms	—	28B
25	Xor 26B-A4B v1.1 Juspay TechnologiesOpen weights	59.4	0.068	110 ms	—	26B

Scores, calibration (ECE) and latency from the Jev Decision Index; cost per 1,000 decisions at 200 input tokens each. Updated daily.

More decision-model rankings

Best decision model Best calibrated Cheapest Runs on CPU Jev alternatives

Frequently asked questions

What is the best open-weights decision model?

Lychee-27B ranks first among open-weights decision models on the Jev Decision Index.

Can open decision models replace a hosted API like Jev?

Several open models now score above Jev on the Decision Index. Check licence (some are non-commercial), size and whether they accept the same request shape.