# DeepSeek V4.1-Flash vs MiniMax M3.1 Flash

## Which cheap, fast model should you use: DeepSeek V4.1-Flash or MiniMax M3.1 Flash?

> It depends. DeepSeek V4.1-Flash scores higher on the Artificial Analysis Intelligence Index (45 vs 43), measured over twice as fast (208.7 vs 93.1 tokens/s), has MIT-licensed open weights and is cheapest off-peak and on cached input. MiniMax M3.1 Flash has flat pricing at any hour, a lower cost per Index task ($0.22 vs $0.26) and a higher HLE score (48% vs 43%), but a 93% hallucination rate on AA-Omniscience. Its weights are only announced for later. Pick DeepSeek for speed, self-hosting and off-peak batch work; pick MiniMax for agent workflows that run all day, with fact checks.

Both are low-cost "Flash" models released in the same week: DeepSeek V4.1-Flash-1017 on October 17 and MiniMax M3.1 Flash on October 19. "Vendor" marks a claim from the model's maker; "AA" marks an independent measurement by Artificial Analysis.

## Side by side

| | DeepSeek V4.1-Flash-1017 | MiniMax M3.1 Flash |
|---|---|---|
| Input, $ per 1M tokens (cache miss) | $0.15 off-peak / $0.30 peak ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)) | $0.15 flat ([MiniMax](https://platform.minimax.io/docs/guides/pricing)) |
| Output, $ per 1M tokens | $0.60 off-peak / $1.20 peak ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)) | $0.60 flat ([MiniMax](https://platform.minimax.io/docs/guides/pricing)) |
| Cached input, $ per 1M tokens | $0.003 off-peak / $0.006 peak ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)) | $0.015 ([MiniMax](https://platform.minimax.io/docs/guides/pricing)) |
| AA Intelligence Index | 45 ([AA](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017)) | 43 on the live page ([AA](https://artificialanalysis.ai/models/minimax-m3-1-flash)); 42 at launch ([VentureBeat](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate)) |
| AA cost per Index task | $0.26 ([AA](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017)) | $0.22 ([AA](https://artificialanalysis.ai/models/minimax-m3-1-flash)) |
| Measured speed (AA) | 208.7 tokens/s, TTFT 1.03 s | 93.1 tokens/s, TTFT 1.39 s |
| Vendor benchmarks | DeepSWE v1.1: 77.0 · Automation-Bench v1.0.6: 60.5 · Agents' Last Exam: 35.4 ([HF card](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017)) | DeepSWE v1.1: 73.7% · Automation-Bench v1.0.6: 62.2 (MiniMax chart, via [VentureBeat](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate)) · Agents' Last Exam: 29.6 ([MiniMax](https://www.minimax.io/blog/minimax-m31f)) |
| Independent tests (AA, via [VentureBeat](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate)) | HLE 43% · CritPt 24% · AA-Omniscience accuracy 43% | HLE 48% · CritPt 22% · AA-Omniscience accuracy 56%, hallucination rate 93% |
| Context window | 1M tokens ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)) | 1,000,000 tokens ([MiniMax](https://platform.minimax.io/docs/guides/text-generation)) |
| Max output | 384K tokens ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)) | Not in official docs |
| Weights / licence | Open weights, MIT ([HF card](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017)) | Not released; announced for later ([MiniMax](https://www.minimax.io/blog/minimax-m31f)) |
| Notable features | Thinking and non-thinking modes, JSON output, tool calls, vision, built-in Search & Fetch APIs | Tunable reasoning effort, interleaved thinking with tool use, text/image/video input, paid web search tool |
| Known limits | Peak hours double the price; very verbose (AA) | Thinking cannot be turned off; high hallucination rate (AA); slower than average (AA) |

## DeepSeek V4.1-Flash-1017

DeepSeek calls this release a "minor update" to V4.1-Flash. The API model name is `deepseek-flash`, and the legacy `deepseek-v4-flash*` names now route to it ([DeepSeek pricing](https://api-docs.deepseek.com/quick_start/pricing), [release news](https://api-docs.deepseek.com/news/news261017)).

- **Peak pricing:** 01:00–04:00 and 06:00–10:00 UTC, Monday to Friday, excluding Chinese public holidays. Off-peak is half price ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)).
- **OpenRouter:** passes through the same time-of-day pricing ([OpenRouter](https://openrouter.ai/api/v1/models)).
- **API features:** thinking mode (default) and non-thinking mode, JSON output, tool calls, Responses API, an Anthropic-format API, chat prefix completion (beta), FIM (beta, non-thinking only) and vision.
- **Concurrency:** limit of 2,500 ([DeepSeek](https://api-docs.deepseek.com/quick_start/pricing)).
- **Caching:** AA lists a 98% cache discount ([AA](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017)).
- **Search & Fetch APIs:** run on DeepSeek's in-house search engine, $0.75 per 1,000 successful requests each. Server-side calls through DeepSeek Harness, Claude Code and Codex are free until billing starts on 2026-10-31 16:00 UTC ([release news](https://api-docs.deepseek.com/news/news261017)).
- **Parameters:** the sources conflict. Hugging Face shows 763B ([HF card](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017)); AA lists 552B total and 16B active ([AA](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017)).
- **Weights:** on Hugging Face and ModelScope under the MIT licence ([release news](https://api-docs.deepseek.com/news/news261017)).

Vendor benchmarks, from the [model card](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017):

- HLE without tools: **42.6** (44.0 on the text-only subset)
- Terminal-Bench 4.0: **42.3**
- ProgramBench: **29.1**
- NL2Repo: **68.2**
- Apex Bench: **72.1**
- DeepSWE cost: DeepSeek claims about **$0.20 per task** at 77.0 ([The Decoder](https://the-decoder.com/deepseek-v4-1-flash-1017-beats-larger-rivals-in-agentic-tasks-while-once-again-massively-cutting-per-task-costs/))

Independent results from [AA](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017):

- Intelligence Index: **45** at "max" reasoning effort
- Output tokens used on the Index: **210M** (AA calls it "very verbose")
- Time to first token: **1.03 s**

## MiniMax M3.1 Flash

MiniMax released M3.1 Flash on October 19. An early version ran anonymously as "Space Bunny" and processed more than 200T tokens ([MiniMax](https://www.minimax.io/blog/minimax-m31f)).

Vendor claims, from the [MiniMax blog](https://www.minimax.io/blog/minimax-m31f):

- About **3×** faster inference and **70%** lower agentic usage cost than its predecessor
- Speed: **200+** tokens/s, P50 **170+**
- BankerToolBench: **76.2**
- OSWorld V2: **61.7**
- CyberGym: **82%**
- #1 among Chinese models on OSWorld V2 and ALE (MiniMax's claim)

Independent results from [AA](https://artificialanalysis.ai/models/minimax-m3-1-flash):

- Intelligence Index: **43** on the live page (42 at launch, per [VentureBeat](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate))
- Output tokens used on the Index: **180M** on the live page (about 189M in the launch figures, per VentureBeat)
- Speed: **93.1** tokens/s, which AA calls "slower than average"
- Time to first token: **1.39 s**
- AA-Omniscience: **56%** accuracy, **93%** hallucination rate (wrong answers as a share of non-correct responses), Omniscience Index **16** (per VentureBeat)

Features and access:

- **Reasoning:** tunable `effort` from low to max (default max). Thinking cannot be turned off; requests that try return a 400 error ([MiniMax docs](https://platform.minimax.io/docs/guides/text-generation)).
- **Tools:** interleaved thinking with tool use ([MiniMax docs](https://platform.minimax.io/docs/guides/text-m3-function-call)).
- **APIs and inputs:** Anthropic- and OpenAI-compatible endpoints; text, image and video input.
- **Paid extras:** `web_search` server tool at $0.01 per request; prompt-cache reads at $0.015 per 1M tokens ([MiniMax pricing](https://platform.minimax.io/docs/guides/pricing)).
- **Where to get it:** MiniMax API, M Plan and MiniMax Code; a Pro model is promised later ([MiniMax](https://www.minimax.io/blog/minimax-m31f)). AA lists one API provider.
- **Weights:** not released yet. MiniMax says they "will be released later" ([MiniMax](https://www.minimax.io/blog/minimax-m31f)); licence terms have not been given ([VentureBeat](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate)). AA marks the model proprietary with no parameter count disclosed.

## Who each suits

**DeepSeek V4.1-Flash** suits teams that want:

- Self-hostable weights under the MIT licence
- The fastest measured API speed of the two
- Built-in web search through the Search & Fetch APIs
- The cheapest cache hits
- Low-cost batch work scheduled off-peak

**MiniMax M3.1 Flash** suits:

- Cost-sensitive agent and automation workloads, such as SaaS workflows and finance tool chains
- Jobs that need the same price at any hour
- Teams willing to add retrieval or verification for anything fact-sensitive, given the 93% hallucination rate

*Facts as of Oct 21, 2026*

## Sources

- [DeepSeek: Models & Pricing](https://api-docs.deepseek.com/quick_start/pricing)
- [DeepSeek: V4.1-Flash-1017 release news](https://api-docs.deepseek.com/news/news261017)
- [Hugging Face: deepseek-ai/DeepSeek-V4.1-Flash-1017 model card](https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017)
- [Artificial Analysis: DeepSeek V4.1-Flash-1017](https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017)
- [MiniMax blog: MiniMax M3.1 Flash](https://www.minimax.io/blog/minimax-m31f)
- [MiniMax: Pricing](https://platform.minimax.io/docs/guides/pricing)
- [MiniMax docs: Text generation](https://platform.minimax.io/docs/guides/text-generation)
- [MiniMax docs: M3 function calling](https://platform.minimax.io/docs/guides/text-m3-function-call)
- [Artificial Analysis: MiniMax M3.1 Flash](https://artificialanalysis.ai/models/minimax-m3-1-flash)
- [VentureBeat: MiniMax's new M3.1 Flash beats DeepSeek and Opus 5 on a key business benchmark at lower cost, but with a high hallucination rate](https://venturebeat.com/technology/minimaxs-new-m3-1-flash-beats-deepseek-and-opus-5-on-a-key-business-benchmark-at-lower-cost-but-with-a-high-hallucination-rate)
- [OpenRouter: models API](https://openrouter.ai/api/v1/models)
- [The Decoder: DeepSeek V4.1-Flash-1017 beats larger rivals in agentic tasks while once again massively cutting per-task costs](https://the-decoder.com/deepseek-v4-1-flash-1017-beats-larger-rivals-in-agentic-tasks-while-once-again-massively-cutting-per-task-costs/)

## Related issues

- [2026-10-18](https://morningtokens.com/issues/2026-10-18/)
- [2026-10-21](https://morningtokens.com/issues/2026-10-21/)

Published Oct 21, 2026