# MiniMax slips a new coding model into its agent tool without a price tag > Source: > Published: 2026-09-27 20:55:10+00:00 *MiniMax launched M3.1-Flash-Preview inside MiniMax Code on September 27, 2026, skipping the usual model card, benchmark report and public API listing entirely. The release lands days after developers fingerprinted a free, anonymous OpenRouter model called Space Bunny Alpha and concluded it's an early build of the same system.* You won't find M3.1-Flash-Preview on OpenRouter. You won't find a price for it anywhere. MiniMax just turned it on inside MiniMax Code, its own coding agent product, and let users start running it. According to MiniMax's own announcement on X from its MiniMax_Agent account, the model "debuts today on MiniMax Code" and is "built for everyday development, fast, reliable, and ready for real work, from quick bug fixes to full features." That's the entire announcement. No model card, no benchmark numbers, no API endpoint developers can call directly. What is public: M3.1-Flash-Preview supports a context window up to 1 million tokens, and MiniMax Code now exposes five reasoning-effort levels for it: low, medium, high, xhigh, and a new tier called max, according to details surfaced by APIMaster.AI and BenchLM.ai. Higher effort levels mean the model thinks longer and produces more output tokens at greater latency, the same trade-off every reasoning model makes, but the tier list itself, five distinct steps instead of the usual three, suggests MiniMax is tuning this specifically for developers who want to dial cost and speed against task difficulty on the fly. Four days before M3.1-Flash-Preview showed up in MiniMax Code, an anonymous model called Space Bunny Alpha appeared on OpenRouter under the identifier stealth/space-bunny-alpha, priced at zero dollars for both input and output during its preview window. Developers didn't wait for MiniMax to say anything. They ran tokenizer comparisons and error-behavior tests, the same fingerprinting techniques the community uses whenever a stealth model shows up trying to get free benchmark data out of the wild. One OpenCode study found a 24 out of 24 token match against MiniMax's known model family; a separate, broader measurement set put it at 50 out of 50, according to reporting from The Neuron and CellCog. Adam Holter, posting on X, stated plainly that Space Bunny Alpha is MiniMax M3.1. MiniMax hasn't confirmed it. But the pattern, free anonymous benchmark run followed days later by a quiet in-product launch of a similarly named model, isn't new, and it's not accidental. It's how you collect real-world performance data without putting your name on a model that might embarrass you. [A free 42x speedup for llama.cpp reveals the real 2026 AI cost lever](https://startupfortune.com/a-free-42x-speedup-for-llamacpp-reveals-the-real-2026-ai-cost-lever/) An open-source technique called prompt lookup decoding is delivering up to 42x speedups in llama.cpp for repetitive tasks like code edits, no new model or hardware required. The catch: a widely cited benchmark found the same technique gives no net speedup on general chat workloads, making it a narrow but real win for anyone self-hosting... - [how to speed up llama.cpp inference costs](https://startupfortune.com/a-free-42x-speedup-for-llamacpp-reveals-the-real-2026-ai-cost-lever/) - [prompt lookup decoding AI optimization technique explained](https://startupfortune.com/a-free-42x-speedup-for-llamacpp-reveals-the-real-2026-ai-cost-lever/) This is the second coding-focused release from MiniMax in recent months. StartupFortune covered the company's earlier coding-only model, M3, which scored 80.5% on SWE-bench Verified and 59.0% on the harder SWE-Bench Pro. VentureBeat reported at the time that M3 beat GPT-5.5 and Gemini 3.1 Pro on that harder benchmark, at what the outlet described as 5 to 10 percent of the cost. M3's official API pricing sits at $0.30 per million input tokens and $1.20 per million output tokens, with cached input priced at $0.06. For comparison, Claude Sonnet 5 runs $2 per million input tokens and $10 per million output, and Claude Opus 5 runs $5 and $25. MiniMax's M2.7, the generation before M3, was already roughly 10 times cheaper than Claude Sonnet on a per-token basis. Frankly, the pricing gap is the whole story here, and MiniMax knows it. A company that can undercut Anthropic and OpenAI by an order of magnitude on input tokens doesn't need flashy benchmark charts to get developers to try the next release. It just needs to ship fast enough that switching costs never get the chance to build up. M3.1-Flash-Preview is MiniMax's third meaningfully distinct coding model release in under a year, and each one has landed while the last one was still being benchmarked by outside labs. The timing isn't isolated to MiniMax, either. Huawei used its annual conference in Shanghai to unveil the Atlas 960 SuperPoD computing cluster, a chip and cluster upgrade arriving just days before a planned Trump-Xi meeting in Washington on September 24, according to wire coverage from the Associated Press carried by ABC News. Huawei's Ascend chips are increasingly the enterprise procurement story underneath China's model race, with GLM-5 and the upcoming DeepSeek V4 both trained and served on domestic silicon rather than Nvidia hardware. As of this month, Kimi K3 leads the broader Chinese model leaderboard with a score of 71.9, ahead of Qwen 3.8 Max at 71.8 and GLM-5.3 at 65.6, per BenchLM.ai's September rankings. None of that context changes what actually happened this week. MiniMax shipped a model without a price, without a benchmark, and without even fully owning up to what it had already put on OpenRouter for free. The strategy isn't secrecy for its own sake. It's iteration speed treated as the product. Western labs publish a model card and a blog post before letting anyone touch a new release. MiniMax is choosing to let the model answer for itself first, and worry about the paperwork later, if at all. **Also read:** [Alibaba's Qwen3.8-27B nearly matches Claude Opus 5.5 motion graphics on a 4090](https://startupfortune.com/alibabas-qwen38-27b-nearly-matches-claude-opus-55-motion-graphics-on-a-4090/) • [An AI math benchmark problem on Apéry-style proofs just got marked solved](https://startupfortune.com/an-ai-math-benchmark-problem-on-apry-style-proofs-just-got-marked-solved/) • [Seaplanes Are Changing How Travelers Reach Palawan's Remote Islands](https://startupfortune.com/seaplanes-are-changing-how-travelers-reach-palawans-remote-islands/) *This article is posted in [Entrepreneurship News](https://startupfortune.com/category/entrepreneurship/), check it out for more related stories.* [Mainland Investors Poured $1.4 Billion Into MiniMax Stock in August](https://startupfortune.com/mainland-investors-poured-14-billion-into-minimax-stock-in-august/) MiniMax Group became mainland China's most bought Hong Kong stock in August 2026, pulling in $1.4 billion via Stock Connect and outpacing buying of Alibaba and Tencent. The rally is backed by real numbers: annualized revenue rocketed from $100 million to $800 million in eight months. - [chinese investors buying minimax ai stock in august](https://startupfortune.com/mainland-investors-poured-14-billion-into-minimax-stock-in-august/) - [mainland china ai startup hong kong stock performance](https://startupfortune.com/mainland-investors-poured-14-billion-into-minimax-stock-in-august/) ## Join the discussion [Open in the community →](https://startupfortune.com/community/) Almost there. Sign in and your reply posts straight away.