Agentic Index
Artificial Analysis's Agentic Index: a composite of the publisher's own agentic and tool-use evaluation runs, on a 0 to 100 scale, as OpenRouter carries it. 94 models ranked, as carried in OpenRouter's catalog.
retrieved sha256 · 0bb53a77713fartificialanalysis.ai ↗
The top 26 of 94, 57.2 down to 35.2 on the Agentic Index. 7 of them are on the frontier below, in the accent. 0 of the 26 columns match the selection below and are drawn white.
Every column stands on zero on the index's own 0 to 100 scale, so a column half as tall is half the score, and the maker's mark hangs under each; the leaderboard below carries every ranked model. Scores are Artificial Analysiss, as OpenRouter carries them.
9 models sit on the frontier: nothing scores higher for less.
- the attractive quadrant
- the frontier
- a priced model, in its maker's tone
- on the frontier, named
- 01Ling 3.0 Tiny4.4 · $0.023 blended
- 02Ling 3.0 Flash19.3 · $0.032 blended · +14.9 at 1.4x
- 03DeepSeek V4 Flash 073141.0 · $0.09 blended · +21.7 at 2.9x
- 04GLM 5.3 Flash50.9 · $0.238 blended · +9.9 at 2.6x
- 05GLM 5.353.1 · $0.402 blended · +2.2 at 1.7x
- 06Muse Spark 1.355.5 · $2.00 blended · +2.4 at 5.0x
- 07Qwen3.8 Max (0902)56.0 · $3.00 blended · +0.5 at 1.5x
- 08Claude Opus 556.5 · $10.00 blended · +0.5 at 3.3x
- 09Claude Fable 5.157.2 · $20.00 blended · +0.7 at 2.0x
93 of 94 ranked models carry a list price; 1 is free or unlisted and not in the field. Price is the blended OpenRouter rate, three input tokens to one output, on a logarithmic axis. Each dot is filled in its maker's tone, as the columns above are. The numbered and named dots are the frontier, cheapest first; each step after the first names what it adds over the step before, and at what multiple of the blended price. A hover over any dot names it and its two figures, a click opens the model, and a model named in the picker above keeps its label on. 14 of 93 priced models sit in the attractive quadrant: at or above the median score of 17.9 and at or below the median blended price of $0.715. The tinted corner is that quadrant. 8 of the 93 crosses match the selection below and are drawn white.
8 of 94.
Artificial Analysis · 0 to 1009 more authors are on this board, the largest with 2 models. Search a name to reach one.
| Position | Model | Author | Score | In / out per M | Context | Listed |
|---|---|---|---|---|---|---|
| 62 | Mistral Medium 3.5 | Mistral | 7.0 | $1.50 / $7.50 | 262k | |
| 70 | Devstral 2 2512 | Mistral | 2.3 | $0.40 / $2.00 | 262k | |
| 71 | Mistral Medium 3.1 | Mistral | 1.9 | $0.44 / $2.20 | 131k | |
| 80 | Ministral 3 14B 2512 | Mistral | 1.1 | $0.20 / $0.20 | 262k | |
| 82 | Mistral Large 3 2512 | Mistral | 1.0 | $0.50 / $1.50 | 262k | |
| 88 | Ministral 3 3B 2512 | Mistral | 0.8 | $0.10 / $0.10 | 131k | |
| 89 | Mistral Small 4 | Mistral | 0.8 | $0.15 / $0.60 | 262k | |
| 91 | Ministral 3 8B 2512 | Mistral | 0.6 | $0.15 / $0.15 | 262k |
The field and the frontier above are the whole board and do not narrow with this table: a frontier drawn through 8 of 94 models would say nothing beats them on both while something in the same board does. The models this table holds are drawn white in the figures instead.
A composite the publisher computes from its own evaluation runs. It orders models on the publisher's tasks, not yours; a gap of a point or two is within what a different task mix would move. It says nothing about price, latency, or how a model behaves on your own cases, which is where Vitna's engine measures quality.
Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Vitna ran none of these evaluations and adjusted none of these figures.
