MiniMax M2.7
Free tier available
- Self-hosted (Free)$0
- API (M2 / M2.5 reference, MiniMax / OpenRouter)$0.30/per 1M input tokens
- API (M2.7)Not yet published
Our pickMiniMax M2.7

StepFun Step 3.5 Flash
Tier-list head-to-head. MiniMax M2.7 takes the A-tier slot — here's the breakdown.
Spec sheet
| Tier | A-tierwin | B-tier |
| Overall score | 8.4 / 10win | 7.8 / 10 |
| Free tier | Yes | Yes |
| Starting price | $0 | $0 |
| Best for | Agentic coding and tool-use workflows on a budget. | Teams building agent systems on Chinese open-weight foundations who want something other than DeepSeek or Q… |
| Last reviewed | 2026-04-27 | 2026-04-17 |
Head-to-head
Rated 1-10 on the same rubric across all 130 tools we cover.
What you'll pay
Look past the headline number -- entry-tier limits drive most cost surprises.
Free tier available
Free tier available
MiniMax-M2.7 (229B total, ~10B active MoE) -- self-evolving agent positioning per vendor benchmarks — StepFun Step 3.5 Flash has no published benchmarks
| Benchmark | Description | Score |
|---|---|---|
| SWE-Bench Pro | 56.22% | |
| Terminal Bench 2 | 57% | |
| SWE Multilingual | 76.5% | |
| Multi SWE Bench | 52.7% | |
| VIBE-Pro | 55.6% |
The decision
Use-case anchors and category strengths, side by side.
Agentic coding and tool-use workflows on a budget. Best price-to-SWE-Bench ratio of any open-weights model in 2026.
Visit MiniMax M2.7Teams building agent systems on Chinese open-weight foundations who want something other than DeepSeek or Qwen, especially if agentic tool-use is the primary workload. Also good for Chinese-market products where StepFun's domestic tuning advantages matter. And for anyone looking to add diversity to their open-weight evaluation matrix beyond the top-3 Chinese labs.
Visit StepFun Step 3.5 FlashBottom line
MiniMax M2.7 edges out StepFun Step 3.5 Flash by 0.6 points (8.4 vs 7.8) -- a A-tier vs B-tier split that's narrow but real. Not a blowout; both belong on a shortlist. The score gap shows up most clearly in the categories that matter for MiniMax M2.7's strengths, so if those categories are your priority, the lead translates.
Pricing-wise, both tools have a free tier (MiniMax M2.7 starts $0, StepFun Step 3.5 Flash starts $0), so you can test either without committing. Compare what each free tier actually unlocks -- usage caps, model access, and feature gates differ a lot more than the headline price suggests, especially as both vendors have tightened limits in 2026.
By use case: pick MiniMax M2.7 when agentic coding and tool-use workflows on a budget. Pick StepFun Step 3.5 Flash when teams building agent systems on chinese open-weight foundations who want something other than deepseek or qwen, especially if agentic tool-use is the primary workload. The two tools aren't fighting for the same person -- they're aiming at adjacent jobs that occasionally overlap. If you're squarely in MiniMax M2.7's lane, the tier-list ranking and the use-case fit point the same direction; if you're in StepFun Step 3.5 Flash's lane, the score gap matters less than the fit.
Bottom line: MiniMax M2.7 is the safer default for most readers, but StepFun Step 3.5 Flash is competitive enough that the tie-breaker is your specific workload, not the spec sheet.
Keep digging
Full MiniMax M2.7 review
Tier A · 8.4/10
Full StepFun Step 3.5 Flash review
Tier B · 7.8/10
MiniMax M2.7 alternatives
Other tools in this lane
StepFun Step 3.5 Flash alternatives
Other tools in this lane
Built from our daily AI-tool sweep, last touched April 27, 2026. Honest tier-list reviews — no affiliate-link pieces disguised as advice. See the rubric or how we review.