Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
C++
{
"available": true,
"version": "v0.1.38",
"repo": "/Applications/Strata/engine",
"summary": "UI preview — simulated model state",
"error": null,
"catalog": {
"families": [
{
"family": "qwen",
"title": "Qwen3.8-Flash-Next",
"about": "the original model",
"experimental": false,
"images": true,
"sizes": [
{
"id": "q2_0",
"model": "Q2_0",
"about": "2-bit, the fastest",
"download_gb": 66.4,
"ram_needed_gb": 48,
"experts_gb": 34,
"on_this_pc": "does not fit",
"installed": false
},
{
"id": "iq2_xs",
"model": "IQ2_XS",
"about": "2-bit i-quant, a little better quality, close in speed",
"download_gb": 68,
"ram_needed_gb": 48,
"experts_gb": 35.5,
"on_this_pc": "does not fit",
"installed": false
},
{
"id": "iq3_xxs",
"model": "IQ3_XXS",
"about": "3-bit i-quant, better quality, slower (more CPU work per token)",
"download_gb": 75.8,
"ram_needed_gb": 60,
"experts_gb": 42.9,
"on_this_pc": "does not fit",
"installed": false
},
{
"id": "iq3_s",
"model": "IQ3_S",
"about": "3.5-bit i-quant, the best quality (matches the full model), the slowest; needs a 64 GB PC with little else running",
"download_gb": 83.6,
"ram_needed_gb": 62,
"experts_gb": 50.3,
"on_this_pc": "does not fit",
"installed": false
}
]
},
{
"family": "swift",
"title": "Swift 1.5",
"about": "thinks much shorter (-63% thinking tokens, 1.8x sooner answers by its authors' numbers)",
"experimental": false,
"images": true,
"sizes": [
{
"id": "swift-iq2_xs",
"model": "IQ2_XS",
"about": "2-bit i-quant, a little better quality, close in speed",
"download_gb": 68,
"ram_needed_gb": 48,
"experts_gb": 35.5,
"on_this_pc": "does not fit",
"installed": false
},
{
"id": "swift-iq3_xxs",
"model": "IQ3_XXS",
"about": "3-bit i-quant, better quality, slower (more CPU work per token)",
"download_gb": 75.8,
"ram_needed_gb": 60,
"experts_gb": 42.9,
"on_this_pc": "does not fit",
"installed": false
}
]
},
{
"family": "coder",
"title": "Qwen3.8-Flash-Next Coder",
"about": "half the experts (code, tools, images kept): needs ~32 GB of RAM, faster; weaker outside coding",
"experimental": false,
"images": true,
"sizes": [
{
"id": "coder-iq1_m",
"model": "IQ1_M",
"about": "the Coder's only size: half the experts, stored like IQ3_S (3.5 bits)",
"download_gb": 58.4,
"ram_needed_gb": 32,
"experts_gb": 23.4,
"on_this_pc": "does not fit",
"installed": false
}
]
},
{
"family": "unsloth",
"title": "Qwen3.8-Flash-Next (Unsloth)",
"about": "4-bit, 111 GB download, most experts read from the SSD: slow (7-8.5 tokens/s on a 64 GB PC)",
"experimental": true,
"images": false,
"sizes": [
{
"id": "unsloth-ud-q4_k_xl",
"model": "UD-Q4_K_XL",
"about": "4-bit (Unsloth Dynamic), EXPERIMENTAL: the best quality, but most experts come from the SSD on a 64 GB PC (7-8.5 tokens/s measured)",
"download_gb": 111.3,
"ram_needed_gb": 48,
"experts_gb": 77,
"on_this_pc": "does not fit",
"installed": false
}
]
}
],
"contexts": [
8192,
32768,
65536,
131072,
262144,
393216,
524288
],
"recommendation": {
"family": "qwen",
"model": "Q2_0",
"context": 32768
},
"context_note": "longer contexts need more VRAM (fewer experts fit on the GPU); past 262144 setup adds rope scaling (experimental)"
},
"fields": [
{
"id": "reasoning_budget_tokens",
"label": "Thinking token budget",
"group": "Responses",
"type": "number",
"min": 0,
"max": 1048576,
"hint": "Leave blank for the model default.",
"supported": true
},
{
"id": "fit_max_tokens",
"label": "Fit output to remaining context",
"group": "Responses",
"type": "boolean",
"hint": "Shortens the requested output cap when the prompt leaves less room.",
"supported": true
},
{
"id": "anthropic_thinking",
"label": "Anthropic thinking",
"group": "Responses",
"type": "enum",
"choices": [
"model",
"on_request"
],
"hint": "Follow the model default or think only when requested.",
"supported": true
},
{
"id": "effort_position",
"label": "Reasoning effort position",
"group": "Responses",
"type": "enum",
"choices": [
"start",
"end"
],
"supported": false
},
{
"id": "aliases",
"label": "API model aliases",
"group": "Serving",
"type": "array",
"hint": "Comma-separated names clients may use for this model.",
"supported": true
},
{
"id": "idle_unload_s",
"label": "Unload after idle seconds",
"group": "Serving",
"type": "number",
"min": 0,
"max": 86400,
"hint": "Zero keeps the model loaded.",
"supported": true
},
{
"id": "lazy_load",
"label": "Load on first request",
"group": "Serving",
"type": "boolean",
"supported": true
},
{
"id": "parallel",
"label": "Concurrent requests",
"group": "Serving",
"type": "number",
"min": 1,
"max": 8,
"hint": "More slots need more memory and can slow each response.",
"supported": false
},
{
"id": "engine_silence_s",
"label": "Engine stall timeout (seconds)",
"group": "Recovery",
"type": "number",
"min": 10,
"max": 3600,
"supported": true
},
{
"id": "api_monitor",
"label": "Keep API request metadata",
"group": "Recovery",
"type": "boolean",
"supported": true
},
{
"id": "layer_split",
"label": "Layer split",
"group": "GPU & memory",
"type": "string",
"hint": "auto, or increasing split points such as 18 or 16,32.",
"supported": true
},
{
"id": "gpu_order",
"label": "GPU order",
"group": "GPU & memory",
"type": "enum",
"choices": [
"auto",
"as_given"
],
"supported": false
},
{
"id": "split_skip_if_fits",
"label": "Use one GPU when the model fits",
"group": "GPU & memory",
"type": "boolean",
"supported": true
},
{
"id": "vram_elastic",
"label": "Elastic GPU expert cache",
"group": "GPU & memory",
"type": "boolean",
"supported": false
},
{
"id": "vram_reserve_mib",
"label": "GPU memory for other apps (MiB)",
"group": "GPU & memory",
"type": "number",
"min": 0,
"max": 131072,
"hint": "Applied as --vram-reserve-mib when loading.",
"supported": true
},
{
"id": "vision_tokens",
"label": "Maximum image tokens",
"group": "Image reading",
"type": "number",
"min": 1,
"max": 16384,
"supported": true
},
{
"id": "vision_min_tokens",
"label": "Minimum image tokens",
"group": "Image reading",
"type": "number",
"min": 0,
"max": 16384,
"supported": true
},
{
"id": "mcp_timeout_s",
"label": "Tool timeout (seconds)",
"group": "Tools",
"type": "number",
"min": 1,
"max": 3600,
"supported": true
},
{
"id": "mcp_max_rounds",
"label": "Maximum tool rounds",
"group": "Tools",
"type": "number",
"min": 1,
"max": 100,
"supported": true
},
{
"id": "mcp_max_result_chars",
"label": "Maximum tool result characters",
"group": "Tools",
"type": "number",
"min": 100,
"max": 1000000,
"supported": true
}
],
"capabilities": {
"updates": false,
"rollback": false,
"calibration": true,
"responses": true
},
"instances": [],
"jobs": [],
"defaultId": null
}