Devstral Small vs Granite-4.0-H-Small vs Qwen Flash
Qwen Flash comes out ahead, 68 to 63 and 57 on our weighted score, though Granite-4.0-H-Small is 17% cheaper per token.
Mistral AIDevstral Small
57/100- ECI—
- Price$0.10 / $0.30
- Context128K
IBMGranite-4.0-H-Small
63/100- ECI—
- Price$0.064 / $0.265
- Context131K
- Our pick
Alibaba (Qwen)Qwen Flash
68/100- ECI—
- Price$0.05 / $0.40
- Context1M
Qwen Flash is our pick
Qwen Flash is the better all-round choice, scoring 68/100 against Granite-4.0-H-Small (63) and Devstral Small (57). It leads on context window. Granite-4.0-H-Small wins on price. 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 priceGranite-4.0-H-SmallGranite-4.0-H-Small $0.114 · Qwen Flash $0.138 · Devstral Small $0.15 per 1M tokens (3:1 blend)
- Longest contextQwen FlashQwen Flash 1,000,000 · Granite-4.0-H-Small 131,072 · Devstral Small 128,000 tokens
- Widest inputsSame inputsDevstral Small: Text · Granite-4.0-H-Small: Text · Qwen Flash: Text
- Self-hostingDevstral Small and Granite-4.0-H-SmallPublishes downloadable weights
| Measure | Weight | Devstral Small | Granite-4.0-H-Small | Qwen Flash |
|---|---|---|---|---|
| Price | 50% | 89 | 95 | 91 |
| Inputs & features | 30% | 25 | 35 | 35 |
| Context window | 20% | 24 | 24 | 60 |
| Overall | 100% | 57/100 | 63/100 | 68/100 |
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 | Devstral SmallMistral AI | Granite-4.0-H-SmallIBM | Qwen FlashAlibaba (Qwen) |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.10 | $0.064 | $0.05 (best) |
| Output | $0.30 | $0.265 (best) | $0.40 |
| Cached input | — | — | — |
| Blended (3:1) | $0.15 | $0.114 (best) | $0.138 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Official Mistral API | Official watsonx.ai API | Official Alibaba API |
| Limits | |||
| Context window | 128,000 tokens | 131,072 tokens | 1,000,000 tokens (best) |
| Max output | 128,000 tokens | 131,072 tokens (best) | 32,768 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | No | No |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | No |
| Reasoning | No | No | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | Open | Open | Proprietary |
| API model ID | devstral-small-2507 | ibm/granite-4-h-small | qwen-flash |
| API providers | 2 | 1 | 6 (best) |
| Released | Jul 10, 2025 | Oct 2, 2025 | Jul 28, 2025 |
| Knowledge cutoff | May 2025 | — | Apr 2024 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- Devstral Small$1.60
- Granite-4.0-H-Small$1.17
- Qwen Flash$1.30
Which should you choose?
Which is better: Devstral Small, Granite-4.0-H-Small or Qwen Flash?
Qwen Flash is the better all-round choice, scoring 68/100 against Granite-4.0-H-Small (63) and Devstral Small (57). It leads on context window. Granite-4.0-H-Small wins on price. 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, Devstral Small, Granite-4.0-H-Small or Qwen Flash?
Granite-4.0-H-Small is cheaper at $0.064 input / $0.265 output per million tokens (official watsonx.ai API price). Qwen Flash costs $0.05 input / $0.40 output per million tokens (official Alibaba API price); Devstral Small costs $0.10 input / $0.30 output per million tokens (official Mistral API price). At a typical mix of three input tokens to one output token, that is $0.114 per million tokens for Granite-4.0-H-Small versus $0.138 for Qwen Flash (1.2× as much) and $0.15 for Devstral Small (1.3× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. Devstral Small has not been scored yet, Granite-4.0-H-Small has not been scored yet and Qwen Flash has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for Devstral Small, Granite-4.0-H-Small and Qwen Flash 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?
Qwen Flash has the largest context window at 1,000,000 tokens, against 131,072 for Granite-4.0-H-Small and 128,000 for Devstral Small. Maximum output per response: Devstral Small up to 128,000, Granite-4.0-H-Small up to 131,072, Qwen Flash up to 32,768 tokens.
Which can read images, PDFs, audio or video?
Devstral Small accepts text; Granite-4.0-H-Small accepts text; Qwen Flash accepts text. They handle the same number of input types.
Are any of these open source?
Devstral Small and Granite-4.0-H-Small publishes its weights and can be self-hosted; Qwen Flash is proprietary.
Which is newer?
Granite-4.0-H-Small is the newest, released Oct 2, 2025. Qwen Flash came out Jul 28, 2025; Devstral Small came out Jul 10, 2025. Knowledge cutoff: Devstral Small May 2025, Qwen Flash Apr 2024.
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.
