Gemini 3.6 Flash vs MiniMax M3

MiniMax M3 has been retired from our catalogue. Its figures below are the last published rates; calls to it are no longer served, while the other model in this comparison is.

vs

Which one, when

Pick gemini-3.6-flash when audio is in the mix: it is the only one of the two that accepts audio input ($0.75 per million audio tokens, $0.075 cached) alongside text, image and video, and its 1048576-token context edges out the 1000000 of the other. minimax-m3 bills $0.3 in and $1.2 out per million against $0.75 and $3.75, so 2.5x cheaper on input and about 3.1x on output, and it allows up to 524288 output tokens where gemini-3.6-flash caps at 65536. minimax-m3 also lets you disable thinking and is flagged for reasoning and long-context work, whereas gemini-3.6-flash always thinks.

Benchmarks

MiniMax M3: the vendor has not published benchmark scores.

Above averageNo peer higherGemini 3.6 Flash6 / 191 / 19
Gemini 3.6 Flash MiniMax M3 other models measured peer average ★ no peer scored higher
DeepSWE 1.1
48.6%
N/A
BioMysteryBench hard
41.2%
N/A
OSWorld 2.0
33.8%
N/A
GDPval-AA v2 Elo · 1422-1598 per Google · 2026-08-25
1422
N/A
Harvey Lab-AA
85.1%
N/A
HLE-Verified
51.2%
N/A
AutomationBench
17%
N/A
LVBench
84.2%
N/A

Vendor-published: Alibaba (Qwen) Alibaba (Qwen) Alibaba (Qwen) Alibaba (Qwen) Anthropic Anthropic Anthropic Anthropic DeepSeek DeepSeek Google Google Google Google Moonshot OpenAI OpenAI OpenAI Tencent Z.ai Z.ai

Pricing

Gemini 3.6 Flash MiniMax M3 Δ
Input / 1M tokens $0.75 $0.3 2.5×
Output / 1M tokens $3.75 $1.2 3.1×
Cache read / 1M tokens $0.075 $0.06 1.3×
Cache write - no separate charge -

Rates from the live catalogue at build time; each model page carries the current rate card.

Where they sit · input price per 1M tokens across all 91 chat models on this billing unit (log scale)

Gemini 3.6 Flash · $0.75 MiniMax M3 · $0.3
$0.05 · Qwen3 VL Flash $30 · GPT-5.4 Pro

Capabilities

Gemini 3.6 Flash MiniMax M3
Tool calling yes yes
Thinking control always on configurable
Structured output yes -
Prompt caching implicit + explicit implicit (automatic)
Cache lifetime not published not published
Minimum cached prefix 4,096 tokens 512 tokens

Specs

Gemini 3.6 Flash MiniMax M3
Input modalities text image audio video text image video
Output modalities text text
Released 2026-07-21 2026-06-01
Knowledge cutoff 2026-03 -
Context window 1,048,576 1,000,000
Max output 65,536 524,288
Thinking parameter thinkingLevel
  • thinking.type
  • reasoning_split
Accepted values
  • minimal
  • low
  • medium
  • high
thinking.type
  • adaptive
  • disabled
reasoning_split
  • boolean
Default medium adaptive: thinking on, with the model deciding when extra reasoning helps

Specs are transcribed from each vendor’s documentation; a row a vendor does not publish is left out rather than inferred. Full sources: Gemini 3.6 Flash · MiniMax M3

One prompt, both models, measured through the gateway

PROMPT Explain to a CFO, in exactly three sentences, why our LLM bill doubled after we enabled reasoning. CHECK exactly 3 sentences

Gemini 3.6 Flash passed · 3 sentences

Enabling reasoning causes the model to generate thousands of internal "thinking" tokens behind the scenes to process logic before delivering the final response. Because our vendor bills us for every token generated—including these intermediate processing steps—our billable output volume per query surged significantly. Since output tokens carry a much higher unit cost than input tokens, this jump in processing volume directly doubled our overall spend.

out 893 tok (+816 thinking) latency 4.7 s

MiniMax M3 passed · 3 sentences

Reasoning models don't just answer — they first generate extensive internal "thinking" tokens to work through the problem step-by-step, and you're billed for all of those tokens even though users never see them. A query that previously cost ~200 tokens may now cost 2,000–10,000 once reasoning is enabled, multiplying per-query spend by roughly 5–50x while the visible output stayed the same size — that's the doubling. The lever is tuning: cap reasoning effort on simple queries, cache the reasoning traces, and reserve the full reasoning model for only your hardest 10–20% of requests, which typically restores 60–80% of the savings.

out 764 tok (+629 thinking) latency 3.5 s

Instruction following (exactly three sentences, so countable), audience shaping (CFO register), and the hidden-thinking billing gap the token meters below expose.

PROMPT This function is supposed to return the longest run of consecutive equal items, but callers report it is off by one on single-element inputs and misses runs that end at the last index. Fix it and explain the bug in one paragraph. CHECK fix passes tests

Gemini 3.6 Flash passed · 8/8 cases

Here is the fixed function along with an explanation of the bugs: ```python def longest_run(xs): if not xs: return 0 best = 1 cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 best = max(best, cur) return best ``` ### Explanation The original function initialized `cur` and `best` to `0` and incremented `cur` on matching adjacent items, which meant it was counting identical adjacent *pairs* rather than the total number of *elements* in a run. This caused single-element inputs to skip the loop and return `0` instead of `1`, and undercounted all run lengths by 1 (for example, a run of two items at the end of the list resulted in `cur = 1`). […]

out 2034 tok (+1768 thinking) latency 10.0 s

MiniMax M3 passed · 8/8 cases

The bug is that the loop starts at index 1 and treats `cur` as the number of *equal transitions* rather than the run length, so a run of *k* consecutive equal items is reported as `k − 1`, and a single-element list (where the loop body never executes) returns 0 instead of 1. A correct version tracks the actual run length by seeding `cur = 1` and resetting it to 1 (not 0) on a mismatch: ```python def longest_run(xs): if not xs: return 0 best = cur = 1 for i in range(1, len(xs)): if xs[i] == xs[i - 1]: cur += 1 else: cur = 1 if cur > best: best = cur return best ``` Now `[5]` returns `1`, `[1,1,1]` returns `3`, and a run like `[1,2,2,2]` that ends at the last index is correctly captured as `3` because `cur` is updated before `best` is compared on every iteration, including the final one.

out 845 tok (+596 thinking) latency 3.5 s

Whether the fix is actually correct (runnable), explanation density, and token efficiency on a bounded task.

PROMPT Extract a JSON object with fields {company, quarter, revenue_usd_m, growth_pct, guidance} from this text. Use null for anything not stated; add no other fields. CHECK valid JSON, schema exact

Gemini 3.6 Flash passed · 5/5 fields, guidance "withheld"

```json { "company": "Meridian Robotics", "quarter": "September quarter", "revenue_usd_m": 412, "growth_pct": null, "guidance": "withheld" } ```

out 2843 tok (+2783 thinking) latency 13.1 s

MiniMax M3 passed · 5/5 fields, guidance null

```json { "company": null, "quarter": "Q3", "revenue_usd_m": 412, "growth_pct": null, "guidance": null } ```

out 1002 tok (+955 thinking) latency 3.6 s

Schema obedience (no invented fields), hallucination pressure (guidance is explicitly withheld), and structured-output path differences.

PROMPT Write a 120-word product announcement for an API gateway feature that caches prompts across providers. Forbidden words: "seamless", "unlock", "game-changer", "revolutionize", "empower". Exactly one sentence must be a question. CHECK 120 words, 0 banned words

Gemini 3.6 Flash missed · 60 words, 0 banned, 0 questions

72: reducing 73: your 74: monthly 75: token 76: spend. S6 (21): 77: You 78: can 79: easily 80: set 81: custom 82: expiration 83: rules, 84: configure 85: TTL 86: settings, 87: and 88: manage 89: cache 90: invalidation 91: across 92: all 93: vendors 94: from 95: one 96: centralized 97: dashboard. S7 (23): 98: Start 99: optimizing […]

out 4092 tok (+3929 thinking) latency 16.3 s

MiniMax M3 passed · 121 words, 0 banned, 1 question

Introducing PromptCache, our intelligent caching layer for the API gateway. Tired of paying twice for identical requests? The feature stores prompt-response pairs across every connected provider, returning cached results instead of hitting costly endpoints. Deployment requires only a single configuration flag, and semantic similarity detection happens automatically. Latency drops up to 70% on repeated traffic while costs follow the same downward curve. Developers access detailed cache hit metrics, eviction logs, and provider breakdowns through the dashboard. Whether routing to OpenAI, Anthropic, or local models, every cached token saves time and money. The cache respects TTL policies, supports manual purging, and encrypts stored content. Available today for Pro and Enterprise tiers—enable PromptCache on existing endpoints and watch inference bills shrink immediately.

out 2808 tok (+2656 thinking) latency 6.0 s

Constraint obedience (word budget, banned-word list, the single question), style fingerprint, and length control.

Switch between them with one line

Both ids are in every tab below; the highlighted pair of lines is the only edit. Same endpoint, same key, same request shape.

from openai import OpenAI

client = OpenAI(
    base_url="https://synthorai.io/v1",
    api_key="sk-syn-...",
)

resp = client.chat.completions.create(
    model="gemini-3.6-flash",
    # model="minimax-m3",  # uncomment this line, comment the one above
    messages=[{"role": "user", "content": "Summarize this diff"}],
)
print(resp.choices[0].message.content)

Get your API key →

FAQ

Which is cheaper, Gemini 3.6 Flash or MiniMax M3?

MiniMax M3 is cheaper on the "Input / 1M tokens" row ($0.3 vs $0.75, 2.5× apart). Other rows may point the other way; the table above carries the full rate card, and real cost depends on your mix.

Can I A/B test Gemini 3.6 Flash against MiniMax M3 without two integrations?

Yes. Both are served through the same OpenAI-compatible endpoint with one API key. Switching is a one-line change to the model id, so you can route a fraction of traffic to each and compare bills directly.

Do Gemini 3.6 Flash and MiniMax M3 support prompt caching?

Yes. Both bill cache reads below their input rate, so warm-prefix workloads cost less than the list rates suggest. The exact cache-read rows are in the pricing table above.

Related comparisons

From our measured studies