DeepSeek-V3 vs GPT-4.1 vs Claude Sonnet 3.5 v2
GPT-4.1 comes out ahead, 65 to 58 and 49 on our weighted score, though DeepSeek-V3 is 6.8× cheaper per token.
Price in rating: Excluded · Include price
DeepSeek
DeepSeek-V3
49/100- ECI131.7
- Price$0.32 / $1.10
- Context131K
- Our pick
OpenAI
GPT-4.1
65/100- ECI136.2
- Price$2.00 / $8.00
- Context1.05M
Anthropic
Claude Sonnet 3.5 v2
58/100- ECI133.3
- Price$2.80 / $14.00
- Context200K
GPT-4.1 is our pick
GPT-4.1 is the better all-round choice, scoring 65/100 against Claude Sonnet 3.5 v2 (58) and DeepSeek-V3 (49). It leads on capability, inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
- CapabilityGPT-4.1Capabilities Index (ECI): GPT-4.1 136.2 · Claude Sonnet 3.5 v2 133.3 · DeepSeek-V3 131.7
- Lowest priceDeepSeek-V3DeepSeek-V3 $0.515 · GPT-4.1 $3.50 · Claude Sonnet 3.5 v2 $5.60 per 1M tokens (3:1 blend)
- Longest contextGPT-4.1GPT-4.1 1,047,576 · Claude Sonnet 3.5 v2 200,000 · DeepSeek-V3 131,072 tokens
- Widest inputsGPT-4.1 and Claude Sonnet 3.5 v2DeepSeek-V3: Text · GPT-4.1: Text, Images, PDFs · Claude Sonnet 3.5 v2: Text, Images, PDFs
- Self-hostingDeepSeek-V3Publishes downloadable weights (DeepSeek Model License)
| Measure | Weight | DeepSeek-V3 | GPT-4.1 | Claude Sonnet 3.5 v2 |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 60 | 64 | 62 |
| Inputs & features | 20% | 25 | 70 | 60 |
| Context window | 13% | 32 | 65 | 39 |
| Overall | 100% | 49/100 | 65/100 | 58/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 131.7 | 136.2 (best) | 133.3 |
| ECI rank | #136 of 158 | #123 of 158 (best) | #131 of 158 |
| GPQA DiamondGraduate-level science questions | 56.5% | 66.9% (best) | 55.3% |
| FrontierMath Tiers 1–3Research-level mathematics | — | 6.0% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 15.8% | 38.3% (best) | 8.5% |
| SWE-bench VerifiedFixing real GitHub issues | — | 48.5% | — |
| SimpleQA VerifiedShort factual questions | — | 31.1% | — |
| Price per million tokens | |||
| Input | $0.32 (best) | $2.00 | $2.80 |
| Output | $1.10 (best) | $8.00 | $14.00 |
| Cached input | — | $0.50 | — |
| Blended (3:1) | $0.515 (best) | $3.50 | $5.60 |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 9 providers | Official OpenAI API | Median of 2 providers |
| Limits | |||
| Context window | 131,072 tokens | 1,047,576 tokens (best) | 200,000 tokens |
| Max output | 8,192 tokens | 32,768 tokens (best) | 8,192 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | Yes | Yes |
| PDFs | No | Yes | Yes |
| Audio | No | No | No |
| Video | No | No | No |
| Reasoning | No | No | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | OpenDeepSeek Model License | Proprietary | Proprietary |
| API model ID | — | gpt-4.1 | — |
| API providers | 9 | 32 (best) | 3 |
| Released | Dec 26, 2024 | Apr 14, 2025 | Oct 22, 2024 |
| Knowledge cutoff | — | Apr 2024 | Apr 30, 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-V3$5.40
GPT-4.1$36.00
Claude Sonnet 3.5 v2$56.00
Which should you choose?
Which is better: DeepSeek-V3, GPT-4.1 or Claude Sonnet 3.5 v2?
GPT-4.1 is the better all-round choice, scoring 65/100 against Claude Sonnet 3.5 v2 (58) and DeepSeek-V3 (49). It leads on capability, inputs & features and context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, DeepSeek-V3, GPT-4.1 or Claude Sonnet 3.5 v2?
DeepSeek-V3 is cheaper at $0.32 input / $1.10 output per million tokens (median across 9 API providers). GPT-4.1 costs $2.00 input / $8.00 output per million tokens (official OpenAI API price); Claude Sonnet 3.5 v2 costs $2.80 input / $14.00 output per million tokens (median across 2 API providers). At a typical mix of three input tokens to one output token, that is $0.515 per million tokens for DeepSeek-V3 versus $3.50 for GPT-4.1 (6.8× as much) and $5.60 for Claude Sonnet 3.5 v2 (11× as much).
Which scores higher on benchmarks?
GPT-4.1 scores higher on the Capabilities Index (ECI): GPT-4.1 136.2 (#123 of 158), Claude Sonnet 3.5 v2 133.3 (#131 of 158) and DeepSeek-V3 131.7 (#136 of 158). The confidence ranges of the top two overlap (134.1–138.1 vs 131.0–135.5), so treat the gap as small. On individual benchmarks: GPQA Diamond — GPT-4.1 66.9%, DeepSeek-V3 56.5%, Claude Sonnet 3.5 v2 55.3%; OTIS Mock AIME 2024–2025 — GPT-4.1 38.3%, DeepSeek-V3 15.8%, Claude Sonnet 3.5 v2 8.5%.
Which is better for coding?
There are no published SWE-bench Verified results for DeepSeek-V3 and Claude Sonnet 3.5 v2 yet, so there is no like-for-like coding score. On overall capability, GPT-4.1 leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
GPT-4.1 has the largest context window at 1,047,576 tokens, against 200,000 for Claude Sonnet 3.5 v2 and 131,072 for DeepSeek-V3. Maximum output per response: DeepSeek-V3 up to 8,192, GPT-4.1 up to 32,768, Claude Sonnet 3.5 v2 up to 8,192 tokens.
Which can read images, PDFs, audio or video?
DeepSeek-V3 accepts text; GPT-4.1 accepts text, images and PDFs; Claude Sonnet 3.5 v2 accepts text, images and PDFs. GPT-4.1 handles the widest range of inputs.
Are any of these open source?
DeepSeek-V3 publishes its weights (DeepSeek Model License) and can be self-hosted; GPT-4.1 and Claude Sonnet 3.5 v2 is proprietary.
Which is newer?
GPT-4.1 is the newest, released Apr 14, 2025. DeepSeek-V3 came out Dec 26, 2024; Claude Sonnet 3.5 v2 came out Oct 22, 2024. Knowledge cutoff: GPT-4.1 Apr 2024, Claude Sonnet 3.5 v2 Apr 30, 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.