GLM-4.6V-Flash vs Llama 3.1 Nemotron Ultra 253B vs Ministral 3 14B
Too close to call on our weighted score (Llama 3.1 Nemotron Ultra 253B 65, GLM-4.6V-Flash 65, Ministral 3 14B 63). The right pick depends on what you value most.
Z.ai (Zhipu)GLM-4.6V-Flash
65/100- ECI—
- Price$0.161 / $0.559
- Context128K
NVIDIALlama 3.1 Nemotron Ultra 253B
65/100- ECI—
- PriceFree / Free
- Context128K
Mistral AIMinistral 3 14B
63/100- ECI—
- Price$0.268 / $0.325
- Context262K
Too close to call
It is close. Our weighted score puts them within a point (Llama 3.1 Nemotron Ultra 253B 65/100, GLM-4.6V-Flash 65/100, Ministral 3 14B 63/100), so choose by what matters most for your work: Llama 3.1 Nemotron Ultra 253B on price and Ministral 3 14B for long inputs. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
- CapabilityNot enough dataNo independent benchmark covers every model here yet
- Lowest priceLlama 3.1 Nemotron Ultra 253BLlama 3.1 Nemotron Ultra 253B Free · GLM-4.6V-Flash $0.261 · Ministral 3 14B $0.282 per 1M tokens (3:1 blend)
- Longest contextMinistral 3 14BMinistral 3 14B 262,144 · GLM-4.6V-Flash 128,000 · Llama 3.1 Nemotron Ultra 253B 128,000 tokens
- Widest inputsGLM-4.6V-FlashGLM-4.6V-Flash: Text, Images, Video · Llama 3.1 Nemotron Ultra 253B: Text · Ministral 3 14B: Text, Images
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | GLM-4.6V-Flash | Llama 3.1 Nemotron Ultra 253B | Ministral 3 14B |
|---|---|---|---|---|
| Price | 50% | 78 | 100 | 76 |
| Inputs & features | 30% | 70 | 35 | 60 |
| Context window | 20% | 24 | 24 | 37 |
| Overall | 100% | 65/100 | 65/100 | 63/100 |
Left out because at least one model lacks the data: capability. The remaining weights were rescaled.
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | GLM-4.6V-FlashZ.ai (Zhipu) | Llama 3.1 Nemotron Ultra 253BNVIDIA | Ministral 3 14BMistral AI |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.161 | Free (best) | $0.268 |
| Output | $0.559 | Free (best) | $0.325 |
| Cached input | — | — | — |
| Blended (3:1) | $0.261 | Free (best) | $0.282 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 2 providers | Official Nvidia API | Median of 2 providers |
| Limits | |||
| Context window | 128,000 tokens | 128,000 tokens | 262,144 tokens (best) |
| Max output | 32,768 tokens | 8,192 tokens | 262,144 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | No | Yes |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | Yes | No | No |
| Reasoning | Yes | Yes | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | No | Yes |
| Availability | |||
| Weights | Open | Open | OpenApache 2.0 |
| API model ID | glm-4.6v-flash | nvidia/llama-3.1-nemotron-ultra-253b-v1 | — |
| API providers | 6 (best) | 1 | 2 |
| Released | Dec 8, 2025 | Apr 7, 2025 | Dec 2, 2025 |
| Knowledge cutoff | — | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- GLM-4.6V-Flash$2.73
- Llama 3.1 Nemotron Ultra 253BFree
- Ministral 3 14B$3.33
Which should you choose?
Which is better: GLM-4.6V-Flash, Llama 3.1 Nemotron Ultra 253B or Ministral 3 14B?
It is close. Our weighted score puts them within a point (Llama 3.1 Nemotron Ultra 253B 65/100, GLM-4.6V-Flash 65/100, Ministral 3 14B 63/100), so choose by what matters most for your work: Llama 3.1 Nemotron Ultra 253B on price and Ministral 3 14B for long inputs. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
Which is cheaper, GLM-4.6V-Flash, Llama 3.1 Nemotron Ultra 253B or Ministral 3 14B?
Llama 3.1 Nemotron Ultra 253B is cheaper at Free input / Free output per million tokens (official Nvidia API price). GLM-4.6V-Flash costs $0.161 input / $0.559 output per million tokens (median across 2 API providers; free on Z.AI); Ministral 3 14B costs $0.268 input / $0.325 output per million tokens (median across 2 API providers). Llama 3.1 Nemotron Ultra 253B is listed as free.
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. GLM-4.6V-Flash has not been scored yet, Llama 3.1 Nemotron Ultra 253B has not been scored yet and Ministral 3 14B has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for GLM-4.6V-Flash, Llama 3.1 Nemotron Ultra 253B and Ministral 3 14B yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.
Which has the bigger context window?
Ministral 3 14B has the largest context window at 262,144 tokens, against 128,000 for GLM-4.6V-Flash and 128,000 for Llama 3.1 Nemotron Ultra 253B. Maximum output per response: GLM-4.6V-Flash up to 32,768, Llama 3.1 Nemotron Ultra 253B up to 8,192, Ministral 3 14B up to 262,144 tokens.
Which can read images, PDFs, audio or video?
GLM-4.6V-Flash accepts text, images and video; Llama 3.1 Nemotron Ultra 253B accepts text; Ministral 3 14B accepts text and images. GLM-4.6V-Flash handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (Apache 2.0), so you can self-host them.
Which is newer?
GLM-4.6V-Flash is the newest, released Dec 8, 2025. Ministral 3 14B came out Dec 2, 2025; Llama 3.1 Nemotron Ultra 253B came out Apr 7, 2025.
How do you decide the winner?
Each model gets a 0–100 score on capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, 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 Oct 4, 2026.
