Qwen3.8 Flash Next vs GPT-6 Luna vs MiniCPM5-2B
GPT-6 Luna comes out ahead, 78 to 70 and 53 on our weighted score, and it is the cheaper option too.
Alibaba (Qwen)Qwen3.8 Flash Next
70/100- ECI—
- Price$0.20 / $0.50
- Context262K
- Our pick
OpenAIGPT-6 Luna
78/100- ECI—
- Price$0.10 / $0.50
- Context1.05M
OpenBMB
MiniCPM5-2B
53/100- ECI—
- Price$0.124 / $0.743
- Context131K
GPT-6 Luna is our pick
GPT-6 Luna is the better all-round choice, scoring 78/100 against Qwen3.8 Flash Next (70) and MiniCPM5-2B (53). It leads on price and context window. 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 priceGPT-6 LunaGPT-6 Luna $0.20 · Qwen3.8 Flash Next $0.275 · MiniCPM5-2B $0.279 per 1M tokens (3:1 blend)
- Longest contextGPT-6 LunaGPT-6 Luna 1,050,000 · Qwen3.8 Flash Next 262,144 · MiniCPM5-2B 131,072 tokens
- Widest inputsQwen3.8 Flash Next and GPT-6 LunaQwen3.8 Flash Next: Text, Images, Video · GPT-6 Luna: Text, Images, PDFs · MiniCPM5-2B: Text
- Self-hostingQwen3.8 Flash Next and MiniCPM5-2BPublishes downloadable weights (qwen-community-1.0 and apache-2.0)
| Measure | Weight | Qwen3.8 Flash Next | GPT-6 Luna | MiniCPM5-2B |
|---|---|---|---|---|
| Price | 50% | 76 | 83 | 76 |
| Inputs & features | 30% | 80 | 80 | 35 |
| Context window | 20% | 37 | 61 | 24 |
| Overall | 100% | 70/100 | 78/100 | 53/100 |
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 | Qwen3.8 Flash NextAlibaba (Qwen) | GPT-6 LunaOpenAI | MiniCPM5-2BOpenBMB |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| GPQA DiamondGraduate-level science questions | — | 90.5% | — |
| FrontierMath Tiers 1–3Research-level mathematics | — | 79.0% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | — | 98.9% | — |
| SimpleQA VerifiedShort factual questions | — | 41.4% | — |
| Price per million tokens | |||
| Input | $0.20 | $0.10 (best) | $0.124 |
| Output | $0.50 (best) | $0.50 (best) | $0.743 |
| Cached input | — | $0.01 | — |
| Blended (3:1) | $0.275 | $0.20 (best) | $0.279 |
| Long-context rate | Same rate | Over 272K: $0.20 / $0.75 | Same rate |
| Price source | Median of 5 providers | Official OpenAI API | Median of 1 providers |
| Limits | |||
| Context window | 262,144 tokens | 1,050,000 tokens (best) | 131,072 tokens |
| Max output | 131,072 tokens (best) | 128,000 tokens | 131,072 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | No |
| PDFs | No | Yes | No |
| Audio | No | No | No |
| Video | Yes | No | No |
| Reasoning | Yes | Yeslow · medium · high · xhigh · max | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | No |
| Availability | |||
| Weights | Openqwen-community-1.0 | Proprietary | Openapache-2.0 |
| API model ID | — | gpt-6-luna | — |
| API providers | 5 | 24 (best) | 1 |
| Released | Aug 27, 2026 | Sep 22, 2026 | Sep 6, 2026 |
| Knowledge cutoff | — | May 18, 2026 | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Qwen3.8 Flash Next$3.00
- GPT-6 Luna$2.00
- MiniCPM5-2B$2.73
Which should you choose?
Which is better: Qwen3.8 Flash Next, GPT-6 Luna or MiniCPM5-2B?
GPT-6 Luna is the better all-round choice, scoring 78/100 against Qwen3.8 Flash Next (70) and MiniCPM5-2B (53). It leads on price and context window. 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, Qwen3.8 Flash Next, GPT-6 Luna or MiniCPM5-2B?
GPT-6 Luna is cheaper at $0.10 input / $0.50 output per million tokens (official OpenAI API price). Qwen3.8 Flash Next costs $0.20 input / $0.50 output per million tokens (median across 5 API providers); MiniCPM5-2B costs $0.124 input / $0.743 output per million tokens (median across 1 API provider). At a typical mix of three input tokens to one output token, that is $0.20 per million tokens for GPT-6 Luna versus $0.275 for Qwen3.8 Flash Next (1.4× as much) and $0.279 for MiniCPM5-2B (1.4× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Qwen3.8 Flash Next has not been scored yet, GPT-6 Luna has not been scored yet and MiniCPM5-2B has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Qwen3.8 Flash Next, GPT-6 Luna and MiniCPM5-2B 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?
GPT-6 Luna has the largest context window at 1,050,000 tokens, against 262,144 for Qwen3.8 Flash Next and 131,072 for MiniCPM5-2B. Maximum output per response: Qwen3.8 Flash Next up to 131,072, GPT-6 Luna up to 128,000, MiniCPM5-2B up to 131,072 tokens.
Which can read images, PDFs, audio or video?
Qwen3.8 Flash Next accepts text, images and video; GPT-6 Luna accepts text, images and PDFs; MiniCPM5-2B accepts text. Qwen3.8 Flash Next handles the widest range of inputs.
Are any of these open source?
Qwen3.8 Flash Next and MiniCPM5-2B publishes its weights (qwen-community-1.0 and apache-2.0) and can be self-hosted; GPT-6 Luna is proprietary.
Which is newer?
GPT-6 Luna is the newest, released Sep 22, 2026. MiniCPM5-2B came out Sep 6, 2026; Qwen3.8 Flash Next came out Aug 27, 2026. Knowledge cutoff: GPT-6 Luna May 18, 2026.
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.
