ERNIE 4.5 21B A3B vs Voxtral Small 24B 2507 vs Qwen Flash
Qwen Flash comes out ahead, 69 to 57 and 55 on our weighted score, though ERNIE 4.5 21B A3B is 11% cheaper per token.
BaiduERNIE 4.5 21B A3B
57/100- ECI—
- Price$0.07 / $0.28
- Context131K
Mistral AIVoxtral Small 24B 2507
55/100- ECI—
- Price$0.10 / $0.40
- Context33K
- Our pick
Alibaba (Qwen)Qwen Flash
69/100- ECI—
- Price$0.05 / $0.40
- Context1M
Qwen Flash is our pick
Qwen Flash is the better all-round choice, scoring 69/100 against ERNIE 4.5 21B A3B (57) and Voxtral Small 24B 2507 (55). It leads on context window. ERNIE 4.5 21B A3B 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 priceERNIE 4.5 21B A3BERNIE 4.5 21B A3B $0.122 · Qwen Flash $0.138 · Voxtral Small 24B 2507 $0.175 per 1M tokens (3:1 blend)
- Longest contextQwen FlashQwen Flash 1,000,000 · ERNIE 4.5 21B A3B 131,072 · Voxtral Small 24B 2507 32,768 tokens
- Widest inputsVoxtral Small 24B 2507ERNIE 4.5 21B A3B: Text · Voxtral Small 24B 2507: Text, Audio · Qwen Flash: Text
- Self-hostingERNIE 4.5 21B A3B and Voxtral Small 24B 2507Publishes downloadable weights (Apache-2.0 and Apache 2.0)
| Measure | Weight | ERNIE 4.5 21B A3B | Voxtral Small 24B 2507 | Qwen Flash |
|---|---|---|---|---|
| Price | 50% | 89 | 82 | 87 |
| Inputs & features | 30% | 25 | 45 | 45 |
| Context window | 20% | 25 | 0 | 60 |
| Overall | 100% | 57/100 | 55/100 | 69/100 |
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 | ERNIE 4.5 21B A3BBaidu | Voxtral Small 24B 2507Mistral AI | Qwen FlashAlibaba (Qwen) |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | — | — | — |
| ECI rank | — | — | — |
| Price per million tokens | |||
| Input | $0.07 | $0.10 | $0.05 (best) |
| Output | $0.28 (best) | $0.40 | $0.40 |
| Cached input | — | — | $0.01 |
| Blended (3:1) | $0.122 (best) | $0.175 | $0.138 |
| Long-context rate | Same rate | Same rate | Over 256K: $0.25 / $2.00 |
| Price source | Median of 1 providers | Official Mistral API | Official Alibaba API |
| Limits | |||
| Context window | 131,072 tokens | 32,768 tokens | 1,000,000 tokens (best) |
| Max output | 8,000 tokens | 32,768 tokens (best) | 32,768 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | No | No |
| PDFs | No | No | No |
| Audio | No | Yes | No |
| Video | No | No | No |
| Reasoning | No | No | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | Yes |
| Availability | |||
| Weights | OpenApache-2.0 | OpenApache 2.0 | Proprietary |
| API model ID | — | voxtral-small-latest | qwen-flash |
| API providers | 1 | 10 (best) | 6 |
| Released | Jun 30, 2025 | Jul 15, 2025 | Jul 28, 2025 |
| Knowledge cutoff | — | — | 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.
- ERNIE 4.5 21B A3B$1.26
- Voxtral Small 24B 2507$1.80
- Qwen Flash$1.30
Which should you choose?
Which is better: ERNIE 4.5 21B A3B, Voxtral Small 24B 2507 or Qwen Flash?
Qwen Flash is the better all-round choice, scoring 69/100 against ERNIE 4.5 21B A3B (57) and Voxtral Small 24B 2507 (55). It leads on context window. ERNIE 4.5 21B A3B 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, ERNIE 4.5 21B A3B, Voxtral Small 24B 2507 or Qwen Flash?
ERNIE 4.5 21B A3B is cheaper at $0.07 input / $0.28 output per million tokens (median across 1 API provider). Qwen Flash costs $0.05 input / $0.40 output per million tokens (official Alibaba API price); Voxtral Small 24B 2507 costs $0.10 input / $0.40 output per million tokens (official Mistral API price). At a typical mix of three input tokens to one output token, that is $0.122 per million tokens for ERNIE 4.5 21B A3B versus $0.138 for Qwen Flash (1.1× as much) and $0.175 for Voxtral Small 24B 2507 (1.4× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. ERNIE 4.5 21B A3B has not been scored yet, Voxtral Small 24B 2507 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 ERNIE 4.5 21B A3B, Voxtral Small 24B 2507 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 ERNIE 4.5 21B A3B and 32,768 for Voxtral Small 24B 2507. Maximum output per response: ERNIE 4.5 21B A3B up to 8,000, Voxtral Small 24B 2507 up to 32,768, Qwen Flash up to 32,768 tokens.
Which can read images, PDFs, audio or video?
ERNIE 4.5 21B A3B accepts text; Voxtral Small 24B 2507 accepts text and audio; Qwen Flash accepts text. Voxtral Small 24B 2507 handles the widest range of inputs.
Are any of these open source?
ERNIE 4.5 21B A3B and Voxtral Small 24B 2507 publishes its weights (Apache-2.0 and Apache 2.0) and can be self-hosted; Qwen Flash is proprietary.
Which is newer?
Qwen Flash is the newest, released Jul 28, 2025. Voxtral Small 24B 2507 came out Jul 15, 2025; ERNIE 4.5 21B A3B came out Jun 30, 2025. Knowledge cutoff: 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 Nov 2, 2026.
