Get started

Epoch Capabilities Index

One number per model, fit by Epoch over the benchmarks that model has results on, with the interval around it. 258 models placed, 116 of them listed in the catalog Vitna prices.

retrieved ·CC-BY-4.0epoch.ai/eci ↗

Models placed258each on 4 benchmark results or more
Results behind it2,917counted here, not republished
Scale130 / 150anchored at Claude 3.5 Sonnet and GPT-5
Released2023 to 2026by the date each model came out

What the interval says

TOP 24 OF 258

Claude Fable 5.2 leads at 173.4, and 8 other models have intervals reaching its own, so the index does not separate them.

  1. 01Claude Fable 5.2Anthropic173.4
  2. 02GPT-6.2 AstraOpenAI172.8
  3. 03Gemini 4 CarbonGoogle DeepMind169.7
  4. 04GPT-6 AstraOpenAI169.2
  5. 05Claude Opus 5.5Anthropic168.7
  6. 06GPT-6.1 SolOpenAI168.6
  7. 07GPT-6.1 TerraOpenAI167.7
  8. 08Claude Fable 5.1Anthropic167.1
  9. 09Gemini 4 HydrogenGoogle DeepMind166.8
  10. 10Claude Sonnet 5.5Anthropic166.6
  11. 11Gemini 4 ArgonGoogle DeepMind166.5
  12. 12Claude Opus 5Anthropic164.7
  13. 13Claude Fable 5Anthropic164.2
  14. 14GPT-6 SolOpenAI163.8
  15. 15GPT-5.6 SolOpenAI163.0
  16. 16GPT-5.5 ProOpenAI162.8
  17. 17Muse WonderMeta AI161.7
  18. 18Kimi K3.1 ProMoonshot161.2
  19. 19GPT-5.6 TerraOpenAI160.8
  20. 20Claude Haiku 5.5Anthropic160.3
  21. 21GPT-5.4 ProOpenAI160.3
  22. 22GPT-5.5OpenAI160.2
  23. 23Claude Opus 4.8Anthropic159.9
  24. 24DeepSeek V4.1 Flash 1017DeepSeek159.4

The rail runs 157.1 to 177.2, the range these 24 rows cover rather than the whole index, because scaled to all 258 they would sit on top of one another. Each bar is the interval Epoch publishes and the mark is the figure; a marked row is one whose interval reaches that of the leader. Epoch states that absolute values are meaningless by themselves and that comparisons between models are what the index is for, that values are not linearly related to benchmark accuracy, and that a highly specialised model may score low while being capable in its own domain.

Vitna fits a scale of its own over the six openly licensed boards, 319 ranked runs over 46 columns, and this is the published index the ordering of that fit is checked against. The two are separate fits over different tables, so their numbers do not compare and only their orderings do.

124 of 258.

Index figures, as Epoch publishes them
#ModelOrganisationWeightsReleasedResultsIntervalIndex
1Claude Fable 5.2AnthropicClosed weights21170.0 to 177.2173.4
2GPT-6.2 AstraOpenAIClosed weights15169.4 to 176.9172.8
3Gemini 4 CarbonGoogle DeepMindClosed weights17166.6 to 173.1169.7
4GPT-6 AstraOpenAIClosed weights26166.4 to 172.5169.2
5Claude Opus 5.5AnthropicClosed weights23165.9 to 172.0168.7
6GPT-6.1 SolOpenAIClosed weights21165.7 to 171.9168.6
7GPT-6.1 TerraOpenAIClosed weights13164.9 to 171.0167.7
8Claude Fable 5.1AnthropicClosed weights25164.4 to 170.1167.1
9Gemini 4 HydrogenGoogle DeepMindClosed weights13164.0 to 170.1166.8
10Claude Sonnet 5.5AnthropicClosed weights20163.9 to 169.6166.6
11Gemini 4 ArgonGoogle DeepMindClosed weights22163.9 to 169.5166.5
12Claude Opus 5AnthropicClosed weights24162.3 to 167.5164.7
13Claude Fable 5AnthropicClosed weights26161.8 to 166.9164.2
14GPT-6 SolOpenAIClosed weights20161.4 to 166.6163.8
15GPT-5.6 SolOpenAIClosed weights27160.7 to 165.6163.0
16GPT-5.5 ProOpenAIClosed weights10160.1 to 165.9162.8
17Muse WonderMeta AIClosed weights13159.4 to 164.3161.7
19GPT-5.6 TerraOpenAIClosed weights20158.5 to 163.3160.8
20Claude Haiku 5.5AnthropicClosed weights20158.2 to 162.7160.3
21GPT-5.4 ProOpenAIClosed weights9157.9 to 163.0160.3
22GPT-5.5OpenAIClosed weights28158.2 to 162.4160.2
23Claude Opus 4.8AnthropicClosed weights26157.8 to 162.2159.9
25Gemini 3.8 FlashGoogle DeepMindClosed weights23156.8 to 161.0158.8
27Gemini 3.7 FlashGoogle DeepMindClosed weights18156.5 to 160.6158.4
28Muse Spark 1.3Meta AIClosed weights14156.2 to 160.5158.2
29GPT-5.4OpenAIClosed weights24156.0 to 160.0157.9
30Grok 4.6xAIClosed weights20155.9 to 160.0157.8
31Qwen 3.8 MaxAlibabaClosed weights13155.5 to 159.8157.6
32Claude Opus 4.7AnthropicClosed weights24155.6 to 159.6157.5
33GPT-5.2 ProOpenAIClosed weights6155.1 to 159.9157.3
36Claude Sonnet 5AnthropicClosed weights18154.9 to 159.1156.9
37GPT-5.6 LunaOpenAIClosed weights21154.9 to 158.8156.7
38GPT-6 LunaOpenAIClosed weights18154.6 to 158.5156.4
39Grok 4.7xAIClosed weights19154.5 to 158.5156.4
40Claude Opus 4.6AnthropicClosed weights27154.6 to 158.3156.3
41Gemini 3.1 ProGoogle DeepMindClosed weights28154.5 to 158.2156.3
42Qwen3.8 Max (0902)AlibabaClosed weights9153.9 to 158.3156.0
45Gemini 3.5 FlashGoogle DeepMindClosed weights18153.5 to 157.3155.3
46Gemini 3.6 FlashGoogle DeepMindClosed weights17153.4 to 157.3155.2
47Muse Spark 1.2Meta AIClosed weights9152.7 to 157.5155.0
48Grok 4.5xAIClosed weights19153.1 to 156.9155.0
49Muse Spark 1.1Meta AIClosed weights7152.4 to 157.7154.9
50Qwen3.7-MaxAlibabaClosed weights12152.6 to 156.8154.6
52GPT-5.2OpenAIClosed weights26152.9 to 156.3154.5
55Muse SparkMeta AIClosed weights5150.9 to 155.5153.2
57Gemini 3 ProGoogle DeepMindClosed weights19151.5 to 154.8153.0
58Claude Sonnet 4.6AnthropicClosed weights21151.2 to 154.5152.8
61Gemini 3 FlashGoogle DeepMindClosed weights20150.8 to 154.1152.3
62Grok 4.20xAIClosed weights15150.5 to 154.1152.3
63Claude Opus 4.5AnthropicClosed weights29150.3 to 153.4151.9
65GPT-5 ProOpenAIClosed weights7149.4 to 153.3151.3
67GPT-5.1OpenAIClosed weights19149.3 to 152.6150.9
72GPT-5OpenAIClosed weights28anchor, fixed150.0
73Grok 4.3 BetaxAIClosed weights11148.0 to 151.9149.9
75GPT-5.4 MiniOpenAIClosed weights15147.7 to 151.1149.4
76Qwen 3.6 PlusAlibabaClosed weights10146.8 to 150.9148.9
80Qwen3.7-PlusAlibabaClosed weights8146.0 to 150.0147.9
82Claude Sonnet 4.5AnthropicClosed weights26146.1 to 149.1147.6
83o3-proOpenAIClosed weights7145.7 to 149.4147.5
85o3OpenAIClosed weights25145.5 to 148.6147.0
87Grok 4xAIClosed weights17145.4 to 148.4146.9
91GPT-5.4 NanoOpenAIClosed weights13144.7 to 148.0146.4
93GPT-5 miniOpenAIClosed weights17144.5 to 147.6146.0
94Gemini 3.5 Flash-LiteGoogle DeepMindClosed weights13144.0 to 147.2145.6
96o4-miniOpenAIClosed weights22143.7 to 146.7145.2
97Gemini 3.1 Flash-LiteGoogleClosed weights8142.8 to 147.1145.0
98Gemini 2.5 Pro (Jun 2025)Google DeepMindClosed weights20143.5 to 146.3144.9
99Claude Opus 4.1AnthropicClosed weights16143.0 to 146.5144.8
101Grok 4 FastxAIClosed weights6142.2 to 146.4144.3
105Gemini 2.5 Pro (Mar 2025)Google DeepMindClosed weights10142.0 to 145.9144.0
110Claude Opus 4AnthropicClosed weights19141.3 to 144.4142.8
111Claude Haiku 4.5AnthropicClosed weights16141.0 to 144.4142.7
112Qwen3-MaxAlibabaClosed weights10140.4 to 144.5142.4
113Gemini 2.5 Flash (Sep 2025)Google DeepMindClosed weights4139.2 to 144.4141.9
114Claude Sonnet 4AnthropicClosed weights21140.3 to 143.3141.8
116Gemini 2.5 Flash (May 2025)Google DeepMindClosed weights7139.4 to 143.6141.6
119o1OpenAIClosed weights17139.5 to 142.8141.2
120Claude 3.7 SonnetAnthropicClosed weights18139.4 to 142.6141.0
122Gemini 2.5 Flash (Jun 2025)Google DeepMindClosed weights7138.3 to 143.1140.8
124Gemini 2.5 Flash (Apr 2025)Google DeepMindClosed weights7138.3 to 142.9140.7
125Gemini 2.5 Pro (May 2025)Google DeepMindClosed weights7138.3 to 143.0140.6
125o3-miniOpenAIClosed weights19138.9 to 142.3140.6
128Grok-3 minixAIClosed weights9138.0 to 141.7139.9
130GPT-5 nanoOpenAIClosed weights17137.1 to 140.9139.1
135Grok 3xAIClosed weights12135.6 to 139.4137.6
137GPT-4.5OpenAIClosed weights11134.4 to 138.4136.5
139GPT-4.1OpenAIClosed weights17134.2 to 138.0136.1
141Gemini 2.0 ProGoogle DeepMindClosed weights5132.6 to 138.0135.4
142o1-miniOpenAIClosed weights10133.0 to 137.3135.3
146GPT-4.1 miniOpenAIClosed weights15132.1 to 136.2134.2
148Gemini 2.0 Flash (Feb 2025)Google DeepMind,GoogleClosed weights11131.7 to 136.2134.0
149Gemini 2.0 Flash Thinking (Jan 2025)Google DeepMind,GoogleClosed weights6130.7 to 136.4133.6
150o1-previewOpenAIClosed weights8131.1 to 135.9133.6
151Claude 3.5 Sonnet (October 2024)AnthropicClosed weights17131.2 to 135.5133.4
152Gemini 2.5 Flash-Lite (Jun 2025)Google DeepMindClosed weights4129.3 to 136.7133.0
153Mistral Medium 3Mistral AIClosed weights5130.1 to 135.3132.8
157Qwen2.5-MaxAlibabaClosed weights6128.4 to 133.7131.2
158Gemini 1.5 Pro (Sept 2024)Google DeepMindClosed weights11127.8 to 132.8130.4
159Claude 3.5 SonnetAnthropicClosed weights12anchor, fixed130.0
160Grok-2 (Dec 2024)xAIClosed weights7127.0 to 132.3129.8
162GPT-4o (Nov 2024)OpenAIClosed weights15126.1 to 131.2128.8
163GPT-4o (May 2024)OpenAIClosed weights8125.4 to 130.7128.1
168GPT-4o (Aug 2024)OpenAIClosed weights10124.1 to 129.9127.2
169Gemini 1.5 Flash (Sep 2024)Google DeepMindClosed weights10123.8 to 129.9127.0
170GPT-4.1 nanoOpenAIClosed weights13123.8 to 129.5126.8
173GPT-4 Turbo (Apr 2024)OpenAIClosed weights8122.5 to 128.8125.8
174Claude 3 OpusAnthropicClosed weights13122.5 to 128.3125.6
175Gemini 1.5 Pro (May 2024)Google DeepMindClosed weights6122.1 to 128.3125.4
176Claude 3.5 HaikuAnthropicClosed weights11121.5 to 127.7124.8
176GPT-4 Turbo (Nov 2023)OpenAIClosed weights6121.5 to 127.7124.8
179GPT-4o miniOpenAIClosed weights19120.5 to 126.8123.8
182GPT-4 (Mar 2023)OpenAIClosed weights7117.6 to 125.2121.6
185Gemini 1.5 Flash (May 2024)Google DeepMindClosed weights6116.7 to 124.1120.6
186Mistral LargeMistral AIClosed weights5115.9 to 123.6120.0
187GPT-4 (Jun 2023)OpenAIClosed weights11116.2 to 123.3119.9
188Claude 3 SonnetAnthropicClosed weights7115.5 to 123.2119.7
192Claude 3 HaikuAnthropicClosed weights9113.5 to 121.5117.8
193Claude 2AnthropicClosed weights6112.8 to 121.2117.2
196Claude InstantAnthropicClosed weights5107.0 to 120.3114.0
201GPT-3.5 Turbo (Jan 2024)OpenAIClosed weights9107.8 to 117.3112.9
202GPT-3.5 Turbo (Nov 2023)OpenAIClosed weights8107.3 to 117.2112.7
210GPT-3.5 Turbo (Jun 2023)OpenAIClosed weights497.6 to 113.4106.0
219Nemotron-4 15BNvidiaClosed weights790.0 to 107.099.1
228Llama 2-34BMeta AIClosed weights883.8 to 102.693.9

124 of 258 models, highest index first. A model links to its catalog page where the pairing names one, which is true of 116; the rest are models the catalog does not list, and the record Epoch keeps of them stands on its own. Results is how many benchmark results stand behind the figure, and Epoch places nothing on fewer than 4. The two rows with no interval are the anchors the scale is defined by, whose figures are fixed rather than estimated.

Sources

Source: the Epoch AI Capabilities Index (epoch.ai/eci), as of 20 Nov 2026. 258 models placed over 2917 benchmark results.
Epoch AI Benchmarking Hub data, by Epoch AI, under CC BY 4.0.
Epoch states that absolute values are meaningless by themselves and that comparisons between models are what the index is for, that values are not linearly related to benchmark accuracy, and that a highly specialised model may score low while being capable in its own domain.
The index is Epoch's fit. Vitna ran none of these evaluations, computed none of these figures and adjusted none of them.