Enterprise AI in Numbers · Capability and cost
Capability gap between the best closed and the best open-weight AI models
11.0 points
11.0 points (Claude Opus 5.5 vs Kimi K3) as of September 22, 2026, according to Epoch AI, Capabilities and benchmarking, up from 10.9 points on September 3, 2026. The best Epoch Capabilities Index score among closed-weight models minus the best among open-weight models released by the date shown.
What it means for leaders
Open-weight models trail the frontier, but by a margin that matters less for many enterprise tasks. Where data control, cost or hosting on your own infrastructure matter more than the last points of capability, open-weight models are a credible option.
What would change the reading: The gap narrowing below the margin between successive closed frontier releases.
The trend
Every reading
| Date | Value | Change | Detail | Source |
|---|---|---|---|---|
| 11.0 points | +0.1 points | Claude Opus 5.5 vs Kimi K3 | Source | |
| 10.9 points | +2.3 points | GPT-6 Astra vs Kimi K3 | Source | |
| 8.6 points | +2.5 points | Claude Fable 5.1 vs Kimi K3 | Source | |
| 6.1 points | +0.5 points | Claude Opus 5 vs Kimi K3 | Source | |
| 5.6 points | −5.9 points | Claude Fable 5 vs Kimi K3 | Source | |
| 11.5 points | −1.0 points | Claude Fable 5 vs GLM-5.2 | Source | |
| 12.5 points | +0.5 points | Claude Fable 5 vs Kimi K2.6 | Source | |
| 12.0 points | +3.4 points | GPT-5.5 Pro vs Kimi K2.6 | Source | |
| 8.6 points | −1.2 points | GPT-5.4 Pro vs Kimi K2.6 | Source | |
| 9.8 points | −1.7 points | GPT-5.4 Pro vs GLM-5.1 | Source | |
| 11.5 points | +2.5 points | GPT-5.4 Pro vs Kimi K2.5 | Source | |
| 9.0 points | +0.7 points | GPT-5.3 Codex vs Kimi K2.5 | Source | |
| 8.3 points | −1.9 points | GPT-5.2 Pro vs Kimi K2.5 | Source | |
| 10.2 points | +3.2 points | GPT-5.2 Pro vs DeepSeek-V3.2 | Source | |
| 7.0 points | −0.3 points | Gemini 3 Pro vs DeepSeek-V3.2 | Source | |
| 7.3 points | +2.5 points | Gemini 3 Pro vs Kimi K2 Thinking | Source | |
| 4.8 points | −0.9 points | GPT-5 Pro vs Kimi K2 Thinking | Source | |
| 5.7 points | +0.6 points | GPT-5 Pro vs DeepSeek-V3.2-Exp | Source | |
| 5.1 points | −1.1 points | GPT-5 vs DeepSeek-V3.2-Exp | Source | |
| 6.2 points | +2.6 points | GPT-5 vs Qwen3-235B-A22B-Thinking (Jul 2025) | Source | |
| 3.6 points | −2.6 points | o3-pro vs Qwen3-235B-A22B-Thinking (Jul 2025) | Source | |
| 6.2 points | +0.6 points | o3-pro vs DeepSeek-R1 (May 2025) | Source | |
| 5.6 points | −2.0 points | o3 vs DeepSeek-R1 (May 2025) | Source | |
| 7.6 points | −0.3 points | o3 vs Qwen3-235B-A22B | Source | |
| 7.9 points | +2.8 points | o3 vs DeepSeek-R1 | Source | |
| 5.1 points | +2.2 points | Gemini 2.5 Pro (Mar 2025) vs DeepSeek-R1 | Source | |
| 2.9 points | −6.5 points | o1 vs DeepSeek-R1 | Source | |
| 9.4 points | −1.8 points | o1 vs DeepSeek-V3 | Source | |
| 11.2 points | +6.0 points | o1 vs Phi-4 | Source | |
| 5.2 points | −1.5 points | o1-mini vs Phi-4 | Source | |
| 6.7 points | −0.2 points | o1-mini vs Qwen2.5-72B | Source | |
| 6.9 points | +5.7 points | o1-mini vs Llama 3.1-405B | Source | |
| 1.2 points | −3.5 points | Claude 3.5 Sonnet vs Llama 3.1-405B | Source | |
| 4.7 points | +1.0 points | Claude 3.5 Sonnet vs Qwen2-72B | Source | |
| 3.7 points | −0.5 points | GPT-4o (May 2024) vs Qwen2-72B | Source | |
| 4.2 points | +1.7 points | GPT-4o (May 2024) vs DeepSeek-V2 (MoE-236B, May 2024) | Source | |
| 2.5 points | −1.8 points | GPT-4 Turbo (Apr 2024) vs DeepSeek-V2 (MoE-236B, May 2024) | Source | |
| 4.3 points | −0.9 points | GPT-4 Turbo (Apr 2024) vs Llama 3-70B | Source | |
| 5.2 points | −3.6 points | GPT-4 Turbo (Apr 2024) vs Mixtral 8x22B | Source | |
| 8.8 points | +0.3 points | GPT-4 Turbo (Apr 2024) vs Mixtral 8x7B | Source | |
| 8.5 points | +0.5 points | Claude 3 Opus vs Mixtral 8x7B | Source | |
| 8.0 points | +0.7 points | GPT-4 Turbo (Nov 2023) vs Mixtral 8x7B | Source | |
| 7.3 points | −1.1 points | GPT-4 (Mar 2023) vs Mixtral 8x7B | Source | |
| 8.4 points | −0.3 points | GPT-4 (Mar 2023) vs Yi-34B | Source | |
| 8.7 points | −3.4 points | GPT-4 (Mar 2023) vs Stable Beluga 2 | Source | |
| 12.1 points | −3.7 points | GPT-4 (Mar 2023) vs Llama 2-70B | Source | |
| 15.8 points | — | GPT-4 (Mar 2023) vs LLaMA-65B | Source |
Definition and method
- What it measures
- The best Epoch Capabilities Index score among closed-weight models minus the best among open-weight models released by the date shown.
- Population
- AI models scored by Epoch AI
- Updated
- When new models are scored
- Source
- Epoch AI, Capabilities and benchmarking · CC BY 4.0
Caveats
- A gap in index points is not a gap in usefulness; for many tasks both are good enough.
- Some models are classified by Epoch as neither open nor closed and are left out.
Cite this
Enterprise AI in Numbers, "Capability gap between the best closed and the best open-weight AI models", https://lucianoayres.com/ai-brief/data/open-weights-gap/, retrieved October 25, 2026. Data: Epoch AI, Capabilities and benchmarking.