Every line, version by version
MiniMax M3.1 Flash vs MiniMax M3: prices per token, context window, benchmark scores, and what changed between the two versions.
Prices are read live from the provider pricing pages, so this table cannot go stale.
| Version | Tier | Status | Released | Input | Output | Cached | Context | Blended /1M | Sample /mo |
|---|---|---|---|---|---|---|---|---|---|
3.1 Released 2026-10-19; $0.15 / $0.60 / $0.015 (¥1 / ¥4 / ¥0.1). Flat rate — no long-context tier, and unlike M3 it does not accept service_tier=priority. Price your own usage | Fast | Active | 2026-10-19 | $0.150 | $0.600 | $0.015 | 1,000K | $0.375 | $0.450 |
3 Released 2026-06-01. service_tier=priority costs 1.5× ($0.45 / $1.80 / $0.09). No separate cache-write charge. The cached-input rate above 512k doubles to $0.12; cost.ts applies the input multiplier to the cached rate as well. Price your own usage | Flagship | Active | 2026-06-01 | $0.300 | $1.20 | $0.060 | 1,000K | $0.750 | $1.80Long context ≥524,288 |
Sample cost = 1M input + 500K output tokens per month at list price. A model with a long-context band is priced on that band — the rule is printed in the row. Use the calculator below for your own workload.
The versions of this line that have a published, traceable score record. 9 benchmarks tracked
MiniMax M3.1 Flash leads on 1 of 1 shared benchmarks; MiniMax M3.1 Flash is ~50% cheaper per blended million tokens.
| Dimension | MiniMax M3.1 Flash | MiniMax M3 |
|---|---|---|
| Terminal-Bench 2.1 | — | 66% |
| SWE-bench Pro | — | 59% |
| DeepSWE v1.1 | 73.7% | — |
| NL2Repo-Bench | — | 42.1% |
| CyberGym | 82% | — |
| Agents' Last Exam | 29.6% | — |
| BrowseComp | — | 83.5% |
| MCP Atlas | — | 74.2% |
| BankerToolBench | 76.2% | 76.1% |
| Blended price | $0.375 /1M | $0.750 /1M |
| Full priceInput · Output · Cached input | $0.150 · $0.600 · $0.015 | $0.300 · $1.20 · $0.060 |
| Context window | 1,000K | 1,000K |
Transcribed from the sources named on each leaderboard entry. Coverage differs per version; a missing cell means the source did not report that benchmark.
Adjust the monthly volume — every version of the line recomputes instantly.
List price only: cached input, batch discounts and off-peak rates are not included. The sample column prices 1M input + 500K output tokens per month.