Vals Web Search Index
Agents work Finance Agent v2 and Legal Research Bench tasks with one web search tool as their only outside source. It compares native provider search with Exa and Keenable.
Domain tools such as SEC EDGAR and CourtListener are removed. Nine model and tool pairs on 2026-10-21, led by GPT-5.6 Sol with Keenable at 49.08%.
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Agent task Many steps in a terminal, browser, desktop, app, game or robot.
GPT-5.6 Sol leads with 49.1% among models released by Jul 2026
Best score per model from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.
The newest model with a score here came out in Jul 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6 Astra
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- and 14 more
- 1 GPT-5.6 Sol 49.1%
- 2 Claude Fable 5 48.5%
- 3
Gemini 3.5 Flash 41.4% - 4 Grok 4.5 38.8%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.6 Sol | 49.1% | Keenable | Jul 2026 | Official leaderboard |
| 2 | Claude Fable 5 | 48.5% | Exa | Jun 2026 | Official leaderboard |
| 3 | Gemini 3.5 Flash | 41.4% | Exa | May 2026 | Official leaderboard |
| 4 | Grok 4.5 | 38.8% | Exa | Jul 2026 | Official leaderboard |
The best score, 49.1%, costs $8.84 a task
Runs on the cost frontier score higher than any run that's cheaper. Claude Fable 5 comes within 0.6 points of the best for $3.48 a task, 3 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-5.6 Sol | Keenable | 49.1% | $8.84 | ||
| Claude Fable 5 | Exa | 48.5% | $3.48 | ||
| Claude Fable 5 | Native | 46.9% | $4.37 | ||
| GPT-5.6 Sol | Exa | 45.2% | $19 | ||
| GPT-5.6 Sol | Native | 43.6% | $0.95 | ||
| Gemini 3.5 Flash | Exa | 41.4% | $1.18 | ||
| Grok 4.5 | Exa | 38.8% | $0.58 | ||
| Gemini 3.5 Flash | Native | 37% | $0.50 | ||
| Grok 4.5 | Native | 35.5% | $0.99 |
2 models have held number one since its release in Jul 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No lab model card reports it yet
We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.
More in Tool use
- AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state. Gemini 4 Carbon 53.9%
- MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks.
Muse Spark 1.1 88.1% - BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes.
Qwen3.7-Max 75% - τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state. Qwen 3.8 Max 55.1%
- Claw-Eval In use 300 human-verified agent tasks (service orchestration, multimodal, multi-turn dialogue) audited via traces and snapshots. Rubric scores with Pass^3 consistency. Claude Sonnet 4.6 81.4%
- DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%