MiniMax M2.5 Pitches Lower Output-Token Costs With Vendor Coding Results
MiniMax released M2.5 and M2.5 Lightning (Feb 12): vendor SWE-Bench Verified 80.2%; Lightning ~100 tok/s at $0.30/$2.40 per MTok. Model weights are on Hugging Face under MiniMax’s published license terms; positioned for continuous agents ~$1/hr in company estimates. Cost-per-task is the trade—not another Elo screenshot.
When a model can sustain high-quality agentic work for roughly $1 per hour, new application classes become practical—persistent agents, always-on monitoring, mass parallel runs. MiniMax’s proprietary frame: sell cost-per-task, not only bench proximity to Opus.
MiniMax launched M2.5 and M2.5 Lightning on February 12. Vendor scores: 80.2% on SWE-Bench Verified; leading or near-leading on Multi-SWE-Bench and BrowseComp (company post). Focus: coding, agentic tool use, search, office productivity; trained with RL across hundreds of thousands of complex environments (company). Lightning: 100 tokens/second (2× typical frontier throughput cited), $0.30/M input and $2.40/M output; standard variant even lower priced; both support caching. MiniMax claims output-token pricing at roughly 1/10–1/20 that of the cited Claude Opus 4.6 and other frontier models. Model weights are available on Hugging Face; the repository metadata says modified MIT, but the current LICENSE-MODEL is a custom MiniMax Model License dated February 13. Full API also offered. Lightning targets speed-critical deployments with identical capability claim vs standard.
Claims vs checks#
Pricing, speed, and license claims are MiniMax primary. SWE-Bench / BrowseComp figures are vendor-reported; an independent leaderboard comparison is not established here. Treat the company’s cost multiples as approximate, not audited TCO. “$1/hour” continuous-agent estimate is company-framed arithmetic.
Limits#
- Cheap tokens ≠ cheap end-to-end agents once tools, retries, and human review enter.
- Review the actual model license and prohibited-use conditions; a repository metadata label is not the full license.
- Throughput claims depend on serving stack and batch size.
Correction and source review — October 14, 2026: Qualified the license claim against the current official model-license text, which is dated February 13; the February 12 launch date remains unchanged. Also clarified that the cost comparison concerns vendor-reported output-token pricing.
Sources#
MiniMax: “MiniMax M2.5: Built for Real-World Productivity” (February 12, 2026). Primary.