Gemma 4 31B IT vs GPT-5.4 nano vs MiniMax-M3
MiniMax-M3 comes out ahead, 73 to 70 and 68 on our weighted score, though Gemma 4 31B IT is 2.5× cheaper per token.
Price in rating: Excluded · Include price
Google
Gemma 4 31B IT
68/100- ECI144.1
- Price$0.142 / $0.40
- Context262K
OpenAI
GPT-5.4 nano
70/100- ECI146.3
- Price$0.20 / $1.25
- Context400K
- Our pick
MiniMax
MiniMax-M3
73/100- ECI147.9
- Price$0.30 / $1.20
- Context1.05M
MiniMax-M3 is our pick
MiniMax-M3 is the better all-round choice, scoring 73/100 against GPT-5.4 nano (70) and Gemma 4 31B IT (68). It leads on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
- CapabilityMiniMax-M3Capabilities Index (ECI): MiniMax-M3 147.9 · GPT-5.4 nano 146.3 · Gemma 4 31B IT 144.1
- Lowest priceGemma 4 31B ITGemma 4 31B IT $0.206 · GPT-5.4 nano $0.463 · MiniMax-M3 $0.525 per 1M tokens (3:1 blend)
- Longest contextMiniMax-M3MiniMax-M3 1,048,576 · GPT-5.4 nano 400,000 · Gemma 4 31B IT 262,144 tokens
- Widest inputsMiniMax-M3Gemma 4 31B IT: Text, Images · GPT-5.4 nano: Text, Images · MiniMax-M3: Text, Images, Video
- Self-hostingGemma 4 31B IT and MiniMax-M3Publishes downloadable weights
| Measure | Weight | Gemma 4 31B IT | GPT-5.4 nano | MiniMax-M3 |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 72 | 74 | 75 |
| Inputs & features | 20% | 70 | 70 | 70 |
| Context window | 13% | 43 | 50 | 65 |
| Overall | 100% | 68/100 | 70/100 | 73/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 144.1 | 146.3 | 147.9 (best) |
| ECI rank | #94 of 158 | #86 of 158 | #77 of 158 (best) |
| GPQA DiamondGraduate-level science questions | 75.8% | 78.5% | 90.9% (best) |
| FrontierMath Tiers 1–3Research-level mathematics | — | 44.9% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 73.3% | 87.8% (best) | 71.1% |
| SimpleQA VerifiedShort factual questions | 10.4% | 11.7% (best) | — |
| Price per million tokens | |||
| Input | $0.142 (best) | $0.20 | $0.30 |
| Output | $0.40 (best) | $1.25 | $1.20 |
| Cached input | — | $0.02 (best) | $0.06 |
| Blended (3:1) | $0.206 (best) | $0.463 | $0.525 |
| Long-context rate | Same rate | Same rate | Over 512K: $0.60 / $2.40 |
| Price source | Median of 42 providers | Official OpenAI API | Official MiniMax (minimax.io) API |
| Limits | |||
| Context window | 262,144 tokens | 400,000 tokens | 1,048,576 tokens (best) |
| Max output | 32,768 tokens | 128,000 tokens | 512,000 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | Yes |
| Reasoning | Yes | Yeslow · medium · high · xhigh | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | No |
| Availability | |||
| Weights | Open | Proprietary | Open |
| API model ID | gemma-4-31b-it | gpt-5.4-nano | MiniMax-M3 |
| API providers | 52 (best) | 36 | 52 (best) |
| Released | Apr 2, 2026 | Mar 17, 2026 | Jun 1, 2026 |
| Knowledge cutoff | Jan 2025 | Aug 31, 2025 | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
Gemma 4 31B IT$2.22
GPT-5.4 nano$4.50
MiniMax-M3$5.40
Which should you choose?
Which is better: Gemma 4 31B IT, GPT-5.4 nano or MiniMax-M3?
MiniMax-M3 is the better all-round choice, scoring 73/100 against GPT-5.4 nano (70) and Gemma 4 31B IT (68). It leads on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, Gemma 4 31B IT, GPT-5.4 nano or MiniMax-M3?
Gemma 4 31B IT is cheaper at $0.142 input / $0.40 output per million tokens (median across 42 API providers; free on Google). GPT-5.4 nano costs $0.20 input / $1.25 output per million tokens (official OpenAI API price); MiniMax-M3 costs $0.30 input / $1.20 output per million tokens (official MiniMax (minimax.io) API price). At a typical mix of three input tokens to one output token, that is $0.206 per million tokens for Gemma 4 31B IT versus $0.463 for GPT-5.4 nano (2.2× as much) and $0.525 for MiniMax-M3 (2.5× as much).
Which scores higher on benchmarks?
MiniMax-M3 scores higher on the Capabilities Index (ECI): MiniMax-M3 147.9 (#77 of 158), GPT-5.4 nano 146.3 (#86 of 158) and Gemma 4 31B IT 144.1 (#94 of 158). The confidence ranges of the top two overlap (145.8–150.0 vs 144.7–148.0), so treat the gap as small. On individual benchmarks: GPQA Diamond — MiniMax-M3 90.9%, GPT-5.4 nano 78.5%, Gemma 4 31B IT 75.8%; OTIS Mock AIME 2024–2025 — GPT-5.4 nano 87.8%, Gemma 4 31B IT 73.3%, MiniMax-M3 71.1%.
Which is better for coding?
There are no published SWE-bench Verified results for Gemma 4 31B IT, GPT-5.4 nano and MiniMax-M3 yet, so there is no like-for-like coding score. On overall capability, MiniMax-M3 leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
MiniMax-M3 has the largest context window at 1,048,576 tokens, against 400,000 for GPT-5.4 nano and 262,144 for Gemma 4 31B IT. Maximum output per response: Gemma 4 31B IT up to 32,768, GPT-5.4 nano up to 128,000, MiniMax-M3 up to 512,000 tokens.
Which can read images, PDFs, audio or video?
Gemma 4 31B IT accepts text and images; GPT-5.4 nano accepts text and images; MiniMax-M3 accepts text, images and video. MiniMax-M3 handles the widest range of inputs.
Are any of these open source?
Gemma 4 31B IT and MiniMax-M3 publishes its weights and can be self-hosted; GPT-5.4 nano is proprietary.
Which is newer?
MiniMax-M3 is the newest, released Jun 1, 2026. Gemma 4 31B IT came out Apr 2, 2026; GPT-5.4 nano came out Mar 17, 2026. Knowledge cutoff: Gemma 4 31B IT Jan 2025, GPT-5.4 nano Aug 31, 2025.
How do you decide the winner?
This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 18, 2026.