Models / Best of /local-32gb
Updated August 2026 · ranked from live data
Best models for 24–32 GB VRAM
Open-weight models that fit a 3090/4090/M-series 32 GB machine at Q4 quantization, newest first.Right now the top pick is Laguna XS 2.1 (~20.3 GB @ Q4).Every row links to full specs, hardware verdicts and provider pricing.
1Laguna XS 2.1poolside~20.3 GB @ Q433.4B$0.060 / $0.12per 1M · via OpenRouter2Nex-N2-MiniNex AGI~22.1 GB @ Q435.1B$0.025 / $0.10per 1M · via OpenRouter3Qwen3.6 27BQwen~17.5 GB @ Q427.8B$0.29 / $2.40per 1M · via OpenRouter4Qwen3.6 35B A3BQwen~22.7 GB @ Q436B / 3B$0.14 / $1.00per 1M · via OpenRouter5Gemma 4 26B A4BGoogle~16.7 GB @ Q426.5B / 4B$0.070 / $0.34per 1M · via OpenRouter6Gemma 4 31BGoogle~20.6 GB @ Q432.7B$0.10 / $0.34per 1M · via OpenRouter7Qwen3.5-35B-A3BQwen~22.7 GB @ Q436B / 3B$0.14 / $1.00per 1M · via OpenRouter8Qwen3.5-27BQwen~17.5 GB @ Q427.8B$0.20 / $1.56per 1M · via OpenRouter9GLM 4.7 FlashZ.AI~19 GB @ Q431.2B$0.060 / $0.40per 1M · via OpenRouter10Nemotron 3 Nano 30B A3BNVIDIA~19.9 GB @ Q431.6B / 3B$0.050 / $0.20per 1M · via Novita AI
