Best closed model
GPT-5.5 (xhigh)
OpenAI
- Intel
- 60.2
- Speed
- 65t/s
- Cost in/out
- €4.31 / €25.86
State-of-the-art reasoning and multimodal tasks requiring maximum intelligence
AI · LEADERBOARD
The AI Leaderboard
Daily ranking of the strongest AI models. No vendor spin, no marketing pages — just the picks teams actually ship with.
Best open model
Kimi K2.6
Kimi
- Intel
- 53.9
- Speed
- 57t/s
- Cost in/out
- Provider€0.82 / €3.45Sovereign~€66 / hr*
High-intelligence open-source tasks with a large parameter footprint
Best value
MiMo-V2-Omni
Xiaomi
- Intel
- 43.4
- Speed
- 108t/s
- Cost in/out
- Sovereign~€1.13 / hr*
Zero-cost value-tier multimodal tasks
Top 10, side by side
| #↑ | Model | Vendor | Intelligence | Speed (tok/s) | Cost (€/Mtok) | Context | Best for |
|---|---|---|---|---|---|---|---|
| 1 | GPT-5.5 (xhigh) | OpenAI | 60.2 | 65 | €4.31 / €25.86 | 922K | State-of-the-art reasoning and multimodal tasks requiring maximum intelligence |
| 2 | GPT-5.5 (high) | OpenAI | 58.9 | 62 | €4.31 / €25.86 | 922K | High-tier enterprise reasoning and complex problem solving |
| 3 | Claude Opus 4.7 (max) | Anthropic | 57.3 | 49 | €5.39 / €21.55 | 1.0M | Deep analytical research and structured extraction over massive contexts |
| 4 | Gemini 3.1 Pro Preview | 57.2 | 138 | €1.72 / €10.35 | 1.0M | High-speed, intelligent multilingual translation and reasoning | |
| 5 | GPT-5.5 (medium) | OpenAI | 56.7 | 59 | €4.31 / €25.86 | 922K | Balanced high-tier reasoning with standard enterprise pricing |
| 6 | Qwen3.7 Max | Alibaba | 56.6 | Hardware dependent | €2.16 / €6.47 | 1.0M | Large-scale multilingual tasks and complex reasoning |
| 7 | Gemini 3.5 Flash | 55.3 | 219 | €1.29 / €7.76 | 1.0M | Ultra-fast, cost-effective customer support and document extraction | |
| 8 | Kimi K2.6 | Kimi | 53.9 | 57 | €0.82 / €3.45 ~€66 / hr* | 256K | High-intelligence open-source tasks with a large parameter footprint |
| 9 | MiMo-V2.5-Pro | Xiaomi | 53.8 | 54 | €0.86 / €2.59 ~€1.13 / hr* | 1.0M | Sovereign, highly efficient open-source deployments with 1M context |
| 10 | GPT-5.3 Codex (xhigh) | OpenAI | 53.6 | 75 | €1.51 / €12.07 | 400K | Advanced code generation and technical reasoning |
Sovereign deploy cost is hourly (Scaleway dedicated GPU cluster, EU, all-in SevenLab pricing). Hover any estimate to see the exact GPU profile. Final price depends on throughput and infra choices.
#1 · proprietary
GPT-5.5 (xhigh)
OpenAI
- Intel
- 60.2
- Speed
- 65 t/s
- Cost
- €4.31 / €25.86
- Ctx
- 922K
State-of-the-art reasoning and multimodal tasks requiring maximum intelligence
#2 · proprietary
GPT-5.5 (high)
OpenAI
- Intel
- 58.9
- Speed
- 62 t/s
- Cost
- €4.31 / €25.86
- Ctx
- 922K
High-tier enterprise reasoning and complex problem solving
#3 · proprietary
Claude Opus 4.7 (max)
Anthropic
- Intel
- 57.3
- Speed
- 49 t/s
- Cost
- €5.39 / €21.55
- Ctx
- 1.0M
Deep analytical research and structured extraction over massive contexts
#4 · proprietary
Gemini 3.1 Pro Preview
- Intel
- 57.2
- Speed
- 138 t/s
- Cost
- €1.72 / €10.35
- Ctx
- 1.0M
High-speed, intelligent multilingual translation and reasoning
#5 · proprietary
GPT-5.5 (medium)
OpenAI
- Intel
- 56.7
- Speed
- 59 t/s
- Cost
- €4.31 / €25.86
- Ctx
- 922K
Balanced high-tier reasoning with standard enterprise pricing
#6 · proprietary
Qwen3.7 Max
Alibaba
- Intel
- 56.6
- Cost
- €2.16 / €6.47
- Ctx
- 1.0M
Large-scale multilingual tasks and complex reasoning
#7 · proprietary
Gemini 3.5 Flash
- Intel
- 55.3
- Speed
- 219 t/s
- Cost
- €1.29 / €7.76
- Ctx
- 1.0M
Ultra-fast, cost-effective customer support and document extraction
#8 · open-source
Kimi K2.6
Kimi
- Intel
- 53.9
- Speed
- 57 t/s
- Cost
- €0.82 / €3.45~€66 / hr*
- Ctx
- 256K
High-intelligence open-source tasks with a large parameter footprint
#9 · open-source
MiMo-V2.5-Pro
Xiaomi
- Intel
- 53.8
- Speed
- 54 t/s
- Cost
- €0.86 / €2.59~€1.13 / hr*
- Ctx
- 1.0M
Sovereign, highly efficient open-source deployments with 1M context
#10 · proprietary
GPT-5.3 Codex (xhigh)
OpenAI
- Intel
- 53.6
- Speed
- 75 t/s
- Cost
- €1.51 / €12.07
- Ctx
- 400K
Advanced code generation and technical reasoning
Which model for which job
Nine scenarios SevenLab teams ship every week. Steal our picks.
High-volume customer support
- Blazing speed of 219 tok/s minimises user latency
- Highly economical pricing at €1.29/Mtok input and €7.76/Mtok output
- Generous 1,000,000 token context handles long chat histories
Gemini 3.5 FlashAdvanced code generation
- Strong intelligence index of 53.56 tailored for coding
- Generous 400,000 token context window for multi-file analysis
- Fast 75 tok/s speed ensures rapid developer feedback loops
GPT-5.3 Codex (xhigh)High-speed document extraction
- Massive 1,000,000 token context window fits entire archives
- Exceptional speed of 219 tok/s accelerates batch processing
- Low cost of €1.29/Mtok input keeps extraction budgets small
Gemini 3.5 FlashDeep long-context analysis
- Massive 1,000,000 token context window for large datasets
- High intelligence index of 57.28 ensures deep analytical accuracy
- Reliable proprietary architecture handles complex reasoning easily
Claude Opus 4.7 (max)Sovereign on-premises deployment
- Strong 53.83 intelligence index beats most open-source rivals
- Compact 7B parameter size minimises local hardware costs
- Massive 1,000,000 token context window under a Mit license
MiMo-V2.5-ProAdvanced multimodal vision
- Unmatched intelligence index of 60.24 for complex visual reasoning
- Large 922,000 token context window handles high-resolution inputs
- Solid 65 tok/s speed ensures responsive multimodal processing
GPT-5.5 (xhigh)Complex agentic automation
- Peak intelligence index of 60.24 ensures reliable tool calling
- Large 922,000 token context window maintains complex agent state
- Closed weights proprietary security protects enterprise workflows
GPT-5.5 (xhigh)High-fidelity multilingual translation
- High intelligence index of 57.18 ensures nuanced translations
- Fast 138 tok/s speed handles high-volume translation tasks
- Massive 1,000,000 token context window translates entire books
Gemini 3.1 Pro PreviewComplex structured extraction
- Massive 1,000,000 token context window handles huge source files
- High intelligence index of 57.28 ensures precise schema adherence
- Reliable proprietary output formatting reduces parsing errors
Claude Opus 4.7 (max)
How we rank
- Snapshot taken 5/22/2026, 2:01:09 AM
- Value score = intelligence index ÷ (input cost + output cost × 3)
Daily refresh at 04:00 Europe/Amsterdam. Numbers move; we don't smooth them.
We pick the model. Then we build the thing.
Building or buying with AI?
Talk directly with our AI specialists
15 min, no strings
No sales pressure
Prototype in 7 days