MiniMax M2.7 vs Qwen3 235B 2507 (Reasoning)

Side-by-side benchmark comparison across agentic, coding, multimodal, knowledge, reasoning, and math workflows.

MiniMax M2.7 is clearly ahead on the aggregate, 57 to 52. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

MiniMax M2.7's sharpest advantage is in agentic, where it averages 57 against 45.9. The single biggest benchmark swing on the page is SWE-bench Pro, 56.2% to 29%.

MiniMax M2.7 is also the more expensive model on tokens at $0.30 input / $1.20 output per 1M tokens, versus $0.00 input / $0.00 output per 1M tokens for Qwen3 235B 2507 (Reasoning). That is roughly Infinityx on output cost alone. Qwen3 235B 2507 (Reasoning) is the reasoning model in the pair, while MiniMax M2.7 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. MiniMax M2.7 gives you the larger context window at 200K, compared with 128K for Qwen3 235B 2507 (Reasoning).

Quick Verdict

Pick MiniMax M2.7 if you want the stronger benchmark profile. Qwen3 235B 2507 (Reasoning) only becomes the better choice if you want the cheaper token bill or you want the stronger reasoning-first profile.

Agentic

MiniMax M2.7

MiniMax M2.7

57

Qwen3 235B 2507 (Reasoning)

45.9

57%
Terminal-Bench 2.0
47%
46.3%
Toolathlon
Coming soon
66.6%
MLE-Bench Lite
Coming soon
62.7%
MM-ClawBench
Coming soon
Coming soon
BrowseComp
48%
Coming soon
OSWorld-Verified
43%

Coding

MiniMax M2.7

MiniMax M2.7

56.2

Qwen3 235B 2507 (Reasoning)

46.2

56.2%
SWE-bench Pro
29%
76.5%
SWE Multilingual
Coming soon
52.7%
Multi-SWE Bench
Coming soon
55.6%
VIBE-Pro
Coming soon
39.8%
NL2Repo
Coming soon
Coming soon
HumanEval
32%
Coming soon
SWE-bench Verified
16%
Coming soon
LiveCodeBench
74.1%

Multimodal & Grounded

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

1495
GDPval-AA
Coming soon
Coming soon
MMMU-Pro
38%
Coming soon
OfficeQA Pro
47%

Reasoning

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

Coming soon
MuSR
36%
Coming soon
BBH
63%
Coming soon
LongBench v2
58%
Coming soon
MRCRv2
58%

Knowledge

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

50
Artificial Analysis
Coming soon
Coming soon
MMLU
40%
Coming soon
GPQA
81.1%
Coming soon
SuperGPQA
64.9%
Coming soon
MMLU-Pro
84.4%
Coming soon
HLE
6%
Coming soon
FrontierScience
42%
Coming soon
SimpleQA
38%

Instruction Following

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

Coming soon
IFEval
87.8%

Multilingual

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

Coming soon
MGSM
62%
Coming soon
MMLU-ProX
81%

Mathematics

Coming soon

Comparable scores for this category are coming soon. One or both models do not have sourced results here yet.

Coming soon
AIME 2023
40%
Coming soon
AIME 2024
42%
Coming soon
AIME 2025
92.3%
Coming soon
HMMT Feb 2023
36%
Coming soon
HMMT Feb 2024
38%
Coming soon
HMMT Feb 2025
37%
Coming soon
BRUMO 2025
39%
Coming soon
MATH-500
60%

Frequently Asked Questions

Which is better, MiniMax M2.7 or Qwen3 235B 2507 (Reasoning)?

MiniMax M2.7 is ahead overall, 57 to 52. The biggest single separator in this matchup is SWE-bench Pro, where the scores are 56.2% and 29%.

Which is better for coding, MiniMax M2.7 or Qwen3 235B 2507 (Reasoning)?

MiniMax M2.7 has the edge for coding in this comparison, averaging 56.2 versus 46.2. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, MiniMax M2.7 or Qwen3 235B 2507 (Reasoning)?

MiniMax M2.7 has the edge for agentic tasks in this comparison, averaging 57 versus 45.9. Inside this category, Terminal-Bench 2.0 is the benchmark that creates the most daylight between them.

Last updated: March 18, 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.