- 🧠 Agentic LLM built for long-horizon software engineering and productivity workflows
- 🆕 First MiniMax model that participates in its own evolution
- 📏 Roughly 198K-token context for extended reasoning and tool use
- 🔧 Function calling, multi-agent "Agent Teams," dynamic tool search
- 🎯 Vendor-reported 56.22% on SWE-Pro, 57.0% on Terminal Bench 2
- 🏢 Built by MiniMax for autonomous productivity and agentic workflows
- 📚 Released March 2026 under a non-commercial MiniMax license
MiniMax is an AI company building generative models across multiple modalities, with a focus that spans both language understanding and audio creation. Their rapid release cadence in early 2026—delivering several new models within just a few months—reflects an ambitious and…
Explore 3 more models by Minimax →MiniMax M2.7 is a text model from MiniMax aimed at autonomous, real-world productivity: complex software engineering, agentic tool use, and office document workflows. Its headline feature is "self-evolution" — MiniMax says an internal version of the model autonomously optimized a programming scaffold over 100+ rounds, analyzing failure trajectories, modifying code, running evaluations, and deciding whether to keep or revert changes, for a reported 30% improvement on internal benchmarks. It also introduces native Agent Teams for multi-agent collaboration with stable role identity and autonomous decision-making.
Within the same family, M2.7 follows [[sibling:minimax-m25|MiniMax M2.5]] and precedes [[sibling:minimax-m3|MiniMax M3]] and its [[sibling:minimax-m3-preview|M3 Preview]]. MiniMax reports significant gains over the previous generation in professional finance tasks — for instance, autonomously reading annual reports and earnings calls, designing assumptions, and building revenue models like a junior analyst.
On vendor-reported evaluations, M2.7 scores 56.22% on SWE-Pro, 76.5 on SWE Multilingual, 55.6% on VIBE-Pro, and 57.0% on Terminal Bench 2, alongside a reported 1495 ELO on GDPval-AA. These figures are self-reported by MiniMax rather than independent evaluators.
The model supports reasoning, code-optimized generation, function calling, and web search, with a context window near 198K tokens. MiniMax describes system-level uses such as correlating monitoring metrics, trace analysis, and SRE-style debugging, citing live incident recovery reduced to under three minutes on multiple occasions. It is distributed under a non-commercial MiniMax license.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
| Seller | Reputation↓ | Input $/M | Cached $/M | Output $/M | Categories | API |
|---|---|---|---|---|---|---|
| Argus AI 0x7adb…c915 | 100 | $0.30 | $0.30 | $1.20 | chat,coding,writing,creative | openai-chat-completions |
| Dark Signal 0x4668…62f2 | 100 | $0.21 | $0.04 | $0.84 | chat,writing,creative | openai-chat-completions |
| Venice.ai Proxy 0x1f22…18c9 | 99 | $0.1875 | $0.0375 | $0.75 | chat,reasoning,coding,web-search | openai-chat-completions |
| Vito-Minimax 0xddfa…27fe | 90 | $0.20 | $0.20 | $1.00 | chat,coding | openai-chat-completions |
| ChainScout AI 0x1734…e621 | 89 | $0.50 | $0.50 | $2.00 | chat,writing,creative,general,model,minimax | openai-chat-completions |
| antseed-zh 0x4122…e194 | 85 | $0.20 | $0.20 | $1.00 | chat,coding | openai-chat-completions |
| Open Ant 0xe4f6…5bc4 | 84 | $0.375 | $0.0688 | $1.50 | chat,tasks,reasoning | openai-chat-completions |
| surplusintelligence.ai 0x0e49…8927 | 79 | $0.1125 | $0.0206 | $0.45 | agents,anon,chat,cheap,code,coding,creative,developer,fast,frontier,function-calling,reasoning,research,tasks,tools,translate,web-search,writing | openai-chat-completions |
| edith 0xb269…b1a6 | 73 | $1.00 | $1.00 | $2.00 | chat,coding | openai-chat-completions |
| BabyCai 0x8509…27b4 | 73 | $0.10 | $0.10 | $0.20 | chat,reasoning | openai-chat-completions |
| ▲ Apex Ant 0x73b4…e736 | 71 | $0.00 | $0.00 | $0.00 | chat,open-source,free,reasoning,long-context,agents | openai-chat-completions |
| Open Bird 0xc0f1…8183 | 70 | $0.105 | $0.021 | $0.42 | chat | openai-chat-completions |
| Fire Ant 🔥🐜 0xbe05…bc5d | 54 | $0.29 | $0.0588 | $1.19 | agent,agents,anon,base-usdc,chat,cheap,code,coding,creative,developer,fast,free,frontier,function-calling,general,github,json,long-context,low-cost,m2-7,m27,math,minimax,model,monitored,open-source,openai-compatible,reasoning,research,response-auth,surplus,tasks,tools,translate,value,verified,web-search,writing | — |
| Night Harbor AI 0x305e…fcf6 | 49 | $0.05 | $0.005 | $0.10 | chat,writing,creative,minimax | openai-chat-completions |
| Meridian AI 0x8c8c…06f5 | 42 | $0.0385 | $0.0385 | $0.154 | chat,reasoning | openai-chat-completions |
| D5V1N2 0xd5e7…7be0 | 41 | $0.18 | $0.036 | $0.72 | chat,coding,tasks,reasoning,cheap,minimax,m27,m2-7 | openai-chat-completions |
| antseed-violet-wren-1ff2 0x7877…1ff2 | 35 | $0.40 | $0.40 | $1.60 | chat,coding,math | openai-chat-completions |
| NovaRoute AI 0xc50d…ed7b | 33 | $0.0216 | $0.0216 | $0.0864 | chat,coding,code,writing,creative,tasks,minimax,value,surplus,openai-compatible,low-cost,verified,github,response-auth,base-usdc,monitored | openai-chat-completions |
| Super Seeder 0xd19f…41f3 | 32 | $0.1875 | $0.0938 | $0.75 | chat,tasks,reasoning | openai-chat-completions |
| uomi.ai 0x87df…48e3 | 30 | $0.223 | $0.223 | $0.96 | chat,math,coding | openai-chat-completions |
| antseed-neon-puma-944e 0x6650…944e | 26 | $0.1875 | $0.0688 | $0.75 | chat,coding,math | openai-chat-completions |
| AntFeed 0xddb6…1442 | 19 | $0.3069 | $0.3069 | $1.2276 | chat | openai-chat-completions |
| Leftermute 0x388b…5389 | 18 | $0.0341 | $0.0341 | $0.1364 | chat,coding,json,tools | openai-chat-completions |
| antseed-tidal-falcon-d92f 0x25e1…d92f | 8 | $0.01 | $0.00 | $0.10 | chat,coding,fast | openai-chat-completions |
| minion0x 0x215e…e2e3 | 5 | $0.13 | $0.03 | $0.58 | chat,coding,math,fast | openai-chat-completions |
"Best price" and the seller table are live AntSeed catalog data (advertised $/1M tokens, not settled amounts). Reputation = on-chain trust (0-100). Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase.