SWE-bench Multimodal (Vals AI) leaderboard
- Claude Opus 5.5 leads
- 70.2%
- models measured
- 18
- lead over #2
- +9.0 pts
- median
- 43.4%
Leader on SWE-bench Multimodal (Vals AI) · read October 21
# SWE-bench Multimodal (Vals AI), read October 21
leaders = {
"Claude Opus 5.5": 70.2%, "Claude Fable 5.1": 61.3%, "Claude Sonnet 5.5": 58.3%,
}
On October 21, 2026, 18 of our catalog's models had a SWE-bench Multimodal (Vals AI) result, and Anthropic Claude Opus 5.5 led with 70.2%, ahead of Anthropic Claude Fable 5.1 at 61.3%.
GitHub issues from visual software, such as front-end and charting libraries, where the issue comes with screenshots. Run by Vals AI. Share resolved.
Role in the index
A coding board we show beside the Vibe Coding Index, not inside it: the index keeps the six boards it was defined on, so a score does not move every time a board is added. Read the methodology
| # | Model | Result | Cost per test | Duration, as printed | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 70.2% | $4.92 | 18m 32s | 88.8 | $20.00 | |
| 2 | Claude Fable 5.1Anthropic | 61.3% | $5.46 | 15m 12s | 84.2 | $50.00 | |
| 3 | Claude Sonnet 5.5Anthropic | 58.3% | $4.53 | 24m 13s | 87.3 | $10.00 | |
| 4 | Gemini 4 ArgonGoogle | 56.0% | $11.40 | 33m 38s | 81.0 | $10.00 | |
| 5 | DeepSeek V4.1 FlashDeepSeek | 46.5% | $0.23 | 19m 24s | 62.5 | $0.78 | |
| 6 | Muse WonderMeta | 45.6% | $1.70 | 9m 44s | 71.9 | $20.00 | |
| 7 | Kimi K3Moonshot AI | 44.2% | $1.65 | 22m 45s | 62.3 | $15.00 | |
| 8 | Claude Haiku 5.5Anthropic | 44.0% | $0.70 | 16m 16s | 70.7 | $0.50 | |
| 9 | GPT-6.1 SolOpenAI | 43.5% | $0.35 | 6m 56s | 82.4 | $10.00 | |
| 10 | GPT-6 AstraOpenAI | 43.3% | $1.81 | 7m 03s | 85.1 | $50.00 | |
| 11 | Grok 4.7xAI | 42.9% | $4.66 | 21m 01s | 66.7 | $6.00 | |
| 12 | Qwen3.8 Max (0902)Qwen | 42.5% | $0.99 | 18m 44s | 62.4 | $6.00 | |
| 13 | MiMo-V2.6-ProXiaomi | 41.3% | $0.12 | 46m 03s | 65.4 | $0.87 | |
| 14 | Gemini 3.8 FlashGoogle | 40.8% | $3.69 | 22m 18s | 61.6 | $3.75 | |
| 15 | Step 5 PreviewStepFun | 40.8% | $0.61 | 26m 06s | 61.1 | $2.70 | |
| 16 | GPT-6 LunaOpenAI | 37.5% | $0.03 | 5m 16s | 57.3 | $0.50 | |
| 17 | MiniMax M3MiniMax | 33.3% | $1.02 | 13m 09s | 39.2 | $1.20 | |
| 18 | Mistral Large 4Mistral AI | 33.1% | $5.14 | 27m 27s | 56.4 | $4.18 |
Only models in our catalog are listed, and each appears once with its best published run. 136 rows were read from the source on October 21, 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.