HybridDeepResearch
Agents combine database queries and web search to answer research questions under constraints.
The paper used the preview split. The release rewrites 88 questions and retains the reference answers and SQL.
- Made by
- Snowflake and collaborators
- Released
- August 2026
- Kind
- Data science and SQL
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Browser, APIs and tools
- Items from
- Real work and use, Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- A version of another benchmark
- In views
- Agents
- Strongest model tested
-
Qwen3.5 397B-A17B #78 on the capability index
- Last model added
- February 2026
- Scores refreshed
- November 20th, 2026
GLM-5.2-FP4 tops the leaderboard with 49.2% Qwen3.5 397B-A17B leads with 32.8% among models released by February 2026
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Qwen3.5 397B-A17B, at 32.8%. This view shows only models we track, each at its best.
The newest model with a score here came out in February 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 GLM-5.2-FP4 49.2%
- 2
Qwen3.5 397B-A17B 32.8%
- 3
Qwen3.5-27B 26.7%
- 4
Qwen3.5-122B-A10B 24.9%
- 5
Qwen3.5-9B 22.3%
- 6
Qwen3.5-4B 17.2%
Source: Official HybridDeepResearch leaderboard
- 1
Qwen3.5 397B-A17B 32.8%
- 2
Qwen3.5-27B 26.7%
- 3
Qwen3.5-122B-A10B 24.9%
- 4
Qwen3.5-9B 22.3%
- 5
Qwen3.5-4B 17.2%
Source: Official HybridDeepResearch leaderboard
| # | Entry | Score | Setting | Agent by | Released | Run on | Source |
|---|---|---|---|---|---|---|---|
| 1 | GLM-5.2-FP4 | 49.2% | MiroFlow | MiroMind | Jun 2026 | Official leaderboard | |
| 2 | | 32.8% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 3 | | 26.7% | smolagents | Hugging Face | Feb 2026 | Jun 2026 | Official leaderboard |
| 4 | | 24.9% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 5 | | 22.3% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 6 | | 17.2% | smolagents | Hugging Face | Feb 2026 | Jun 2026 | Official leaderboard |
| # | Model | Score | Setting | Agent by | Released | Run on | Source |
|---|---|---|---|---|---|---|---|
| 1 | | 32.8% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 2 | | 26.7% | smolagents | Hugging Face | Feb 2026 | Jun 2026 | Official leaderboard |
| 3 | | 24.9% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 4 | | 22.3% | MiroFlow | MiroMind | Feb 2026 | Jun 2026 | Official leaderboard |
| 5 | | 17.2% | smolagents | Hugging Face | Feb 2026 | Jun 2026 | Official leaderboard |
Rankings by part
We read 3 parts from the benchmark’s own leaderboard: Search to SQL, SQL to search and Parallel. Each ranks models on one part of the test. Qwen3.5 397B-A17B leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
Search to SQL
5 models we track have a score here.
- 1
Qwen3.5 397B-A17B 22.2%
- 2
Qwen3.5-122B-A10B ↑1 18.5%
- 3
Qwen3.5-27B ↓1 17.2%
- 4
Qwen3.5-9B 4.96%
- 5
Qwen3.5-4B 3.23%
SQL to search
5 models we track have a score here.
- 1
Qwen3.5 397B-A17B 28.4%
- 2
Qwen3.5-122B-A10B ↑1 22.7%
- 3
Qwen3.5-27B ↓1 21.1%
- 4
Qwen3.5-9B 17.3%
- 5
Qwen3.5-4B 11%
Parallel
5 models we track have a score here.
- 1
Qwen3.5 397B-A17B 60.2%
- 2
Qwen3.5-27B 51.9%
- 3
Qwen3.5-9B ↑1 45.8%
- 4
Qwen3.5-4B ↑1 39.4%
- 5
Qwen3.5-122B-A10B ↓2 37.3%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official HybridDeepResearch leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official HybridDeepResearch leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Coding
- SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
- LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests.
DeepSeek-V4-Pro 93.5%
- SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
- DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
- SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases.
Seed2.1 Pro 59.8%
- Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance.
DeepSeek V4.1 Flash 3,471
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (6 entries).
- Pass Models scored 5 models we track, plus 1 other entry on its leaderboard.
- Warning Newest score The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3.5 397B-A17B (#78).
- Pass Headroom The best is 32.8%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on snowflake-ai-research.github.io: 10 rows for models we track and 1 other entry at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.