FITMYLLM · NOVEMBER 4, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
250.29B
Parameters (15B active)
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 256K 512K 1M
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 75.5 GB
74.9 + 0.6 KV
low IQ2_M 2.93 92.7 GB
92.2 + 0.6 KV
low Q2_K 3.16 99.9 GB
99.4 + 0.6 KV
low IQ3_XXS 3.25 102.7 GB
102.2 + 0.6 KV
low IQ3_XS 3.5 110.6 GB
110.0 + 0.6 KV
low Q3_K_S 3.64 114.9 GB
114.4 + 0.6 KV
low IQ3_M 3.76 118.7 GB
118.1 + 0.6 KV
low Q3_K_M 4 126.2 GB
125.6 + 0.6 KV
low Q3_K_L 4.3 135.6 GB
135.0 + 0.6 KV
moderate IQ4_XS 4.46 140.6 GB
140.0 + 0.6 KV
moderate Q4_K_S 4.67 147.2 GB
146.6 + 0.6 KV
moderate Q4_K_M 4.89 154.0 GB
153.5 + 0.6 KV
good Q5_K_S 5.57 175.3 GB
174.8 + 0.6 KV
good Q5_K_M 5.7 179.4 GB
178.8 + 0.6 KV
good Q6_K 6.56 206.3 GB
205.7 + 0.6 KV
excellent Q8_0 8.5 267.0 GB
266.4 + 0.6 KV
lossless FP16 16 501.6 GB
501.1 + 0.6 KV
lossless
Select your GPU above to see speed estimates and compatibility for each quantization.
Too big for a single GPU — plan a multi-GPU deployment
Even the lightest quant needs ~76 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →
▸ READY TO RUN THIS? RENT BY THE HOUR
RENT A GPU AND RUN SOLAR OPEN 2 250B NOW
Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.
Community Ratings Chat Coding Reasoning Creative Vision Roleplay Agentic
Loading ratings...
Run this model IQ2_XXS — 74.9 GB VRAM IQ2_M — 92.2 GB VRAM Q2_K — 99.4 GB VRAM IQ3_XXS — 102.2 GB VRAM IQ3_XS — 110.0 GB VRAM Q3_K_S — 114.4 GB VRAM IQ3_M — 118.1 GB VRAM Q3_K_M — 125.6 GB VRAM Q3_K_L — 135.0 GB VRAM IQ4_XS — 140.0 GB VRAM Q4_K_S — 146.6 GB VRAM Q4_K_M — 153.5 GB VRAM Q5_K_S — 174.8 GB VRAM Q5_K_M — 178.8 GB VRAM Q6_K — 205.7 GB VRAM Q8_0 — 266.4 GB VRAM FP16 — 501.1 GB VRAM
Ollama llama.cpp vLLM LM Studio KoboldCpp Jan Docker
▸ Easiest way to get started · Beginners
DOCS ↗ curl -fsSL https://ollama.com/install.sh | shCOPY
$ ollama run solar:250b-q4_K_MCOPY
Tag may need adjustment — check ollama.com/library/solar for available tags.
▸ SETUP GUIDE >_
Auto-setup with fitmyllm CLI Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.
Auto-detect GPU Live tok/s in chat Speed benchmarks 9 inference engines
GPUs that can run this model At Q4_K_M quantization. Sorted by minimum VRAM.
Find the best GPU for Solar Open 2 250B
Build Hardware for Solar Open 2 250B ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
Solar Open 2 250B — 250.29B MoE. ▸ SPECIFICATIONS
PARAMETERS 250.29B (15B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 1024K tokens
CAPABILITIES chat, coding, reasoning, multilingual, math, agentic, tool_use
RELEASE DATE 2026-07-22
PROVIDER Upstage
FAMILY solar ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 74.9 GB 65% IQ2_M 2.93 92.2 GB 75% Q2_K 3.16 99.4 GB 78% IQ3_XXS 3.25 102.2 GB 82% IQ3_XS 3.5 110.0 GB 84% Q3_K_S 3.64 114.4 GB 85% IQ3_M 3.76 118.1 GB 86% Q3_K_M 4 125.6 GB 88% Q3_K_L 4.3 135.0 GB 90% IQ4_XS 4.46 140.0 GB 92% Q4_K_S 4.67 146.6 GB 93% Q4_K_M 4.89 153.5 GB 94% Q5_K_S 5.57 174.8 GB 96% Q5_K_M 5.7 178.8 GB 96% Q6_K 6.56 205.7 GB 97% Q8_0 8.5 266.4 GB 100% FP16 16 501.1 GB 100%
§ 01 BENCHMARK SCORES
MMLU-PRO 86.2
GPQA Diamond 85.7
LiveCodeBench 92.4
AIME 95.7
HLE 28.5
aa_scicode 48.0
aa_lcr 71.7
SWE-bench 70.4
aa_terminal_bench_v2 44.2
§ 03 COMPATIBLE GPUs
13 @ Q4_K_M Feedback