Qwen4 Max on Together AI: 1M context digested 652K-token Tornado repo, $0.69 per fix
togethercompute · x · 2026-11-22
Together AI's blog details Qwen4 Max (Alibaba's flagship) on its Serverless Inference with a 1M-token context window, 30% off through Nov 22.
- Test: one request with a bug report plus the entire Tornado repo (344 files, 652,519 tokens), no tools or retrieval. In 72 seconds it returned a patch passing the maintainers' tests, and even found a precedent elsewhere in the repo. Cost: $0.69
- Sale pricing: $1.05/M input tokens, $0.0525/M cached input, $3.15/M output
- Caveats: requests must stream, and thinking is on by default with thinking tokens billed as output
Related event: Qwen4 Max Launches on Together AI with 1M-Token Context(2 posts)→
More from coding & agent
- Firecrawl Alexandria indexes 4M+ podcast episodes for AI agents, with weekly updates — devdigest · 2026-11-22
- Solo consultant's 11-tool AI stack: Claude Code builds, Codex reviews, bots run marketing — jacob_posel · 2026-11-22
- Turn your AI edits into a reusable writing style guide with one prompt — hwchase17 · 2026-11-22
- Neon agent turns photos into Blender 3D models via colmap in one hour — eigenhector · 2026-11-22
- Letter of Agency: Signed, Expiring Credentials to Stop Agents Getting Blocked — moss_in_it · 2026-11-22
- Gemini 4 Neon autonomously writes GPU kernels, 5.9x faster prefill on AMD hardware — sebkrier · 2026-11-22