DeepSeek V4 Flash 0731 vs Llama 4 Scout 17B Instruct
DeepSeek V4 Flash 0731 comes out ahead, 70 to 61 on our weighted score, and it is the cheaper option too.
Price in rating: Excluded · Include price
- Our pick
DeepSeek
DeepSeek V4 Flash 0731
70/100- ECI154.4
- Price$0.14 / $0.30
- Context1M
Meta
Llama 4 Scout 17B Instruct
61/100- ECI127.8
- Price$0.17 / $0.59
- Context10M
Add a model
Make it a three-way comparison.
DeepSeek V4 Flash 0731 is our pick
DeepSeek V4 Flash 0731 is the better all-round choice, scoring 70/100 against Llama 4 Scout 17B Instruct (61). It leads on capability. Llama 4 Scout 17B Instruct wins on inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
- CapabilityDeepSeek V4 Flash 0731Capabilities Index (ECI): DeepSeek V4 Flash 0731 154.4 · Llama 4 Scout 17B Instruct 127.8
- Lowest priceDeepSeek V4 Flash 0731DeepSeek V4 Flash 0731 $0.18 · Llama 4 Scout 17B Instruct $0.275 per 1M tokens (3:1 blend)
- Longest contextLlama 4 Scout 17B InstructLlama 4 Scout 17B Instruct 10,000,000 · DeepSeek V4 Flash 0731 1,000,000 tokens
- Widest inputsLlama 4 Scout 17B InstructDeepSeek V4 Flash 0731: Text · Llama 4 Scout 17B Instruct: Text, Images
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | DeepSeek V4 Flash 0731 | Llama 4 Scout 17B Instruct |
|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 81 | 56 |
| Inputs & features | 20% | 35 | 50 |
| Context window | 13% | 64 | 100 |
| Overall | 100% | 70/100 | 61/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | ||
|---|---|---|
| Capability | ||
| Capabilities Index (ECI) | 154.4 (best) | 127.8 |
| ECI rank | #52 of 158 (best) | #140 of 158 |
| GPQA DiamondGraduate-level science questions | 91.0% (best) | 51.8% |
| FrontierMath Tiers 1–3Research-level mathematics | 57.5% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 94.4% (best) | 7.8% |
| SimpleQA VerifiedShort factual questions | 33.6% | — |
| Price per million tokens | ||
| Input | $0.14 (best) | $0.17 |
| Output | $0.30 (best) | $0.59 |
| Cached input | — | — |
| Blended (3:1) | $0.18 (best) | $0.275 |
| Long-context rate | Same rate | Same rate |
| Price source | Median of 65 providers | Median of 13 providers |
| Limits | ||
| Context window | 1,000,000 tokens | 10,000,000 tokens (best) |
| Max output | 384,000 tokens (best) | 16,384 tokens |
| Inputs and features | ||
| Text | Yes | Yes |
| Images | No | Yes |
| PDFs | No | No |
| Audio | No | No |
| Video | No | No |
| Reasoning | Yes | No |
| Tool calling | Yes | Yes |
| Structured output | No | No |
| Availability | ||
| Weights | OpenMIT | Open |
| API model ID | — | — |
| API providers | 66 (best) | 13 |
| Released | Jul 31, 2026 | Apr 5, 2025 |
| Knowledge cutoff | May 2025 | Aug 2024 |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
DeepSeek V4 Flash 0731$2.00
Llama 4 Scout 17B Instruct$2.88
Which should you choose?
Which is better: DeepSeek V4 Flash 0731 or Llama 4 Scout 17B Instruct?
DeepSeek V4 Flash 0731 is the better all-round choice, scoring 70/100 against Llama 4 Scout 17B Instruct (61). It leads on capability. Llama 4 Scout 17B Instruct wins on inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, DeepSeek V4 Flash 0731 or Llama 4 Scout 17B Instruct?
DeepSeek V4 Flash 0731 is cheaper at $0.14 input / $0.30 output per million tokens (median across 65 API providers). Llama 4 Scout 17B Instruct costs $0.17 input / $0.59 output per million tokens (median across 13 API providers). At a typical mix of three input tokens to one output token, that is $0.18 per million tokens for DeepSeek V4 Flash 0731 versus $0.275 for Llama 4 Scout 17B Instruct (1.5× as much).
Which scores higher on benchmarks?
DeepSeek V4 Flash 0731 scores higher on the Capabilities Index (ECI): DeepSeek V4 Flash 0731 154.4 (#52 of 158) and Llama 4 Scout 17B Instruct 127.8 (#140 of 158). Their confidence ranges do not overlap (152.6–156.4 vs 124.9–130.4), so the gap is a real one. On individual benchmarks: GPQA Diamond — DeepSeek V4 Flash 0731 91.0%, Llama 4 Scout 17B Instruct 51.8%; OTIS Mock AIME 2024–2025 — DeepSeek V4 Flash 0731 94.4%, Llama 4 Scout 17B Instruct 7.8%.
Which is better for coding?
There are no published SWE-bench Verified results for DeepSeek V4 Flash 0731 and Llama 4 Scout 17B Instruct yet, so there is no like-for-like coding score. On overall capability, DeepSeek V4 Flash 0731 leads, which tends to carry over to coding, but test on your own codebase. Both support tool calling for agent workflows.
Which has the bigger context window?
Llama 4 Scout 17B Instruct has the largest context window at 10,000,000 tokens, against 1,000,000 for DeepSeek V4 Flash 0731. Maximum output per response: DeepSeek V4 Flash 0731 up to 384,000, Llama 4 Scout 17B Instruct up to 16,384 tokens.
Which can read images, PDFs, audio or video?
DeepSeek V4 Flash 0731 accepts text; Llama 4 Scout 17B Instruct accepts text and images. Llama 4 Scout 17B Instruct handles the widest range of inputs.
Are any of these open source?
Yes, both publish their weights (MIT), so you can self-host them.
Which is newer?
DeepSeek V4 Flash 0731 is the newest, released Jul 31, 2026. Llama 4 Scout 17B Instruct came out Apr 5, 2025. Knowledge cutoff: DeepSeek V4 Flash 0731 May 2025, Llama 4 Scout 17B Instruct Aug 2024.
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.