Updated 2026-10-07
On JANCTION Render a 1080p Cycles still of a product or interior scene takes 5 to 10 GPU-seconds (0.5 to 1 JPY at the planned 0.1 JPY per GPU-second), a 24-frame 720p turntable MP4 takes 58 GPU-seconds (about 6 JPY), and a one-minute 24 fps video lands between about 150 JPY (EEVEE, 720p) and 1,200 JPY (Cycles, 1080p, 128 samples), estimated from the measured per-frame times below. During the free beta nothing is charged: every key gets 10 GPU-minutes a day, up to 240 frames at 1080p per job. Prices are planned, not yet in force (as of 2026-10-07).
All rows are real renders on the production GPU (NVIDIA RTX PRO 6000 Blackwell, 96 GB), the same records as the benchmarks and examples. Price = GPU seconds × 0.1 JPY (planned rate). During the beta the charge is 0.
| Render | Settings | GPU seconds | At the planned rate |
|---|---|---|---|
| 4-frame preview (tiled sheet) | 640x360 tiles, 16 samples | 4.0 | 0.4 JPY |
| Still: extruded metal logo, sunset | 1920x1080, Cycles, 128 samples | 6.2 | 0.6 JPY |
| Still: living room from primitives | 1600x900, Cycles, 128 samples | 5.4 | 0.5 JPY |
| Still: office towers at night | 1920x1080, Cycles, 128 samples | 9.6 | 1.0 JPY |
| Still: glass bottle (caustics, glass) | 1440x1440, Cycles, 128 samples | 166.5 | 16.7 JPY |
| Turntable: headphones, 24 frames, MP4 | 1280x720, Cycles, 64 samples | 58.3 | 5.8 JPY |
| Clip: wave through 144 cubes, 48 frames, MP4 | 1280x720, EEVEE, 64 samples | 56.6 | 5.7 JPY |
| Clip: letters rising, 48 frames, MP4 | 1280x720, EEVEE, 64 samples | 45.9 | 4.6 JPY |
| Look | Per frame (measured) | GPU seconds for 1,440 frames | At the planned rate |
|---|---|---|---|
| EEVEE, 720p, 64 samples (simple geometry) | about 1.0 s | about 1,450 | about 150 JPY |
| Cycles, 720p, 64 samples (turntable of a product) | about 2.4 s | about 3,500 | about 350 JPY |
| Cycles, 1080p, 128 samples (product or interior still per frame) | 6 to 10 s | about 8,600 to 14,400 | about 860 to 1,440 JPY |
Your scene will differ: glass, volumes and high sample counts cost more (the bottle above is 20 times the logo). The honest way to know is to run render_estimate on your own scene, which answers from the same measurements. During the beta a job is at most 240 frames, so a one-minute video is six jobs.
| Service | Their published price | Notes |
|---|---|---|
| BlendSwap Render Farm | 43 credits per GPU hour, about $4.30 (L40S) | checked 2026-09-30; .blend or zip only; REST API and MCP |
| GarageFarm.NET | GPU: $0.004 / $0.006 / $0.012 per OctaneBench-hour (low / medium / high priority); CPU: $0.024 / $0.036 / $0.072 per GHz-hour; $25 trial credits | jobs go through the renderBeamer desktop client and plugins; no agent API |
| Fox Renderfarm | CPU from $1.224 per node-hour (tiers down to $0.734 with larger recharges); GPU prices behind a toggle not readable on the page; $25 free trial | online submission or desktop client |
| RebusFarm | pricing page not reachable on 2026-10-07 (HTTP 404) | see their site |
| JANCTION Render | free beta: 10 GPU-minutes per key per day; planned 0.1 JPY per GPU-second, prepaid; planned monthly plans from 2,980 JPY for 15 GPU-hours | MCP, HTTP API, CLI; bpy scripts and 3D files accepted |
Units differ (GPU-hour, OctaneBench-hour, GHz-hour, node-hour), so the only like-for-like comparison is your own scene on each service. The farms above are built for large jobs on many nodes; JANCTION Render is built for an agent that previews, fixes and renders in minutes on one datacenter GPU. See the comparison and pricing.
| from | how |
|---|---|
| A file on this computer (.blend, 3D file) | Drop it at /upload to get a 12-hour link (no sign-up), then tell the AI "Render this file: <link>". From inside Blender use the add-on; from a terminal the CLI uploads directly |
| Claude.ai (web / desktop / mobile) | Settings → Connectors → Add custom connector → https://render.janction.jp/mcp. "Connect" opens a page that creates a free key |
| ChatGPT | Settings → Connectors → Advanced → Developer mode → Create → https://render.janction.jp/mcp (OAuth) |
| Claude Code | claude mcp add --transport http janction-render https://render.janction.jp/mcp, then /mcp to authenticate |
| Codex | codex mcp add janction-render --url https://render.janction.jp/mcp |
| Cursor / VS Code | Add to Cursor Add to VS Code The button opens the editor; confirm and it is added. Manual: Streamable HTTP at https://render.janction.jp/mcp |
| Gemini CLI / Antigravity CLI | gemini extensions install https://github.com/JasmyLab-JANCTION/janction-render |
| Grok / Perplexity / Le Chat | Add a custom (remote MCP) connector with https://render.janction.jp/mcp in each app's Connectors settings (Grok: New Connector → Custom; Perplexity: Pro / Max / Enterprise; Le Chat: workspace admin) |
| Windsurf, Cline, Goose, other MCP clients | Streamable HTTP at https://render.janction.jp/mcp; OAuth or Authorization: Bearer <api key> |
| Terminal (stdio, sends local files) | claude mcp add janction-render -e JANCTION_RENDER_SERVER=https://render.janction.jp -- uvx --from janction-render janction-render-mcp |
| Inside Blender (for people) | The add-on; extension repository https://render.janction.jp/extensions/index.json. Maya / Houdini / Cinema 4D: submit tools |
No. The beta is free: 10 GPU-minutes per key per day, jobs up to 240 frames at 1080p. The 0.1 JPY per GPU-second rate is planned and will be announced before it starts.
Call render_estimate (MCP) or POST /v1/estimate (HTTP) with the same body as the render. It returns GPU seconds, a sentence, and whether it fits today's quota.
Glass with refraction and caustics needs far more samples per pixel to converge; the logo is opaque metal under an HDRI. Sample count, resolution and materials drive the time, not the frame count alone.
For clips of a few dozen frames or more, yes (about half per frame on simple scenes). Each EEVEE job pays about 7 seconds of shader compilation, so single stills and previews stay on Cycles.
JANCTION Render v0.4.20 · JasmyLab Inc. · Updated 2026-10-07 · Tested with Blender 5.0, JANCTION Render 0.4.20 · 日本語