Vals Multimodal Index leaderboard
- Claude Fable 5 leads
- 74.2%
- models measured
- 33
- lead over #2
- +0.2 pts
- median
- 60.6%
Leader on Vals Multimodal Index · read October 22
On October 22, 2026, 33 of our catalog's models had a Vals Multimodal Index result, and Anthropic Claude Fable 5 led with 74.2%, ahead of Anthropic Claude Opus 5 at 73.9%.
Finance, coding and education tasks that need images, combined by Vals AI. An archived board.
Role in the index
Vals Multimodal Index is shown as published and does not feed any of our indexes. Read the methodology
| # | Model | Result | Cost per test | Duration, as printed | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Fable 5Anthropic | 74.2% | $7.97 | 12m 49s | 79.3 | $50.00 | |
| 2 | Claude Opus 5Anthropic | 73.9% | $6.14 | 14m 14s | 83.0 | $25.00 | |
| 3 | Kimi K3Moonshot AI | 73.4% | $1.70 | 15m 06s | 62.3 | $15.00 | |
| 4 | GPT-5.6 SolOpenAI | 72.6% | $5.36 | 9m 11s | 73.5 | $20.00 | |
| 5 | Claude Opus 4.8Anthropic | 70.9% | $5.46 | 16m 20s | 65.0 | $25.00 | |
| 6 | Muse Spark 1.2Meta | 69.8% | $0.51 | 7m 26s | 55.7 | $4.25 | |
| 7 | GPT-5.6 LunaOpenAI | 69.1% | $0.16 | 6m 47s | 58.0 | $1.20 | |
| 8 | Claude Sonnet 5Anthropic | 68.8% | $6.53 | 16m 52s | 57.7 | $10.00 | |
| 9 | GPT-5.5OpenAI | 68.1% | $3.32 | 8m 24s | 71.3 | $30.00 | |
| 10 | Claude Opus 4.7Anthropic | 67.4% | $4.44 | 9m 24s | 67.5 | $25.00 | |
| 11 | Muse Spark 1.1Meta | 66.7% | $0.36 | 4m 48s | 70.3 | $4.25 | |
| 12 | Qwen3.8 Max (0902)Qwen | 65.4% | $1.93 | 33m 16s | 62.4 | $6.00 | |
| 13 | Gemini 3.6 FlashGoogle | 65.1% | $1.02 | 8m 56s | 62.5 | $3.75 | |
| 14 | GPT-5.6 TerraOpenAI | 65.1% | $0.80 | 4m 18s | 63.9 | $12.00 | |
| 15 | Grok 4.5xAI | 63.4% | $0.91 | 5m 19s | 58.6 | $6.00 | |
| 16 | Gemini 3.5 FlashGoogle | 62.9% | $1.01 | 4m 26s | 39.8 | $9.00 | |
| 17 | Claude Sonnet 4.6Anthropic | 60.6% | $1.58 | 8m 56s | 57.3 | $15.00 | |
| 18 | MiniMax M3MiniMax | 60.0% | $1.08 | 17m 56s | 39.2 | $1.20 | |
| 19 | Kimi K2.6Moonshot AI | 56.4% | $0.52 | 15m 04s | 52.1 | $4.00 | |
| 20 | Gemini 3.1 Pro PreviewGoogle | 56.1% | $0.98 | 5m 23s | 32.3 | $12.00 | |
| 21 | GPT-5.4 MiniOpenAI | 54.2% | $0.49 | 13m 08s | 35.8 | $4.50 | |
| 22 | Qwen3.7 PlusQwen | 53.9% | $0.26 | 10m 29s | 33.2 | $1.28 | |
| 23 | MiMo-V2.5Xiaomi | 52.8% | $0.03 | 7m 10s | 50.0 | $0.336 | |
| 24 | Gemini 3 Flash PreviewGoogle | 52.2% | $0.29 | 4m 20s | 43.7 | $3.00 | |
| 25 | Qwen3.6 PlusQwen | 51.5% | $1.12 | 10m 43s | 44.1 | $1.95 | |
| 26 | T Inkling SmallThinking Machines | 50.0% | $0.14 | 11m 49s | 27.3 | $1.20 | |
| 27 | T InklingThinking Machines | 49.4% | $0.63 | 15m 29s | 31.5 | $4.05 | |
| 28 | GPT-5.4 NanoOpenAI | 47.6% | $0.23 | 11m 11s | 44.8 | $1.25 | |
| 29 | Grok 4.3xAI | 43.3% | $0.50 | 5m 39s | 37.4 | $2.50 | |
| 30 | Claude Haiku 4.5Anthropic | 42.9% | $0.41 | 4m 43s | 35.6 | $5.00 | |
| 31 | Gemini 3.1 Flash Lite PreviewGoogle | 41.4% | $0.12 | 2m 32s | 22.7 | $1.50 | |
| 32 | Grok 4.20xAI | 39.1% | $0.40 | 2m 51s | 33.3 | $2.50 | |
| 33 | Mistral Medium 3.5Mistral AI | 34.8% | $3.45 | 12m 15s | 25.5 | $7.50 |
Only models in our catalog are listed, and each appears once with its best published run. 137 rows were read from the source on October 22, 2026, and rows that match no catalog model are left out. Cost per test is what Vals AI reports one test cost on the run behind the result, in US dollars. Every other column is a figure the board publishes for the same run.