Benchmark Almanac

PaperScope

Grok 4 41.0 leads since January 2026
Not in model cards

Agents search multiple scientific papers to compare figures, summarize results, and solve problems. Each question can require evidence from several documents.

Made by
Lei Xiong and colleagues
Released
January 2026
Kind
Deep research
Level
Expert
Scored
Graded by a model
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Documents, Tables, Images → Text, Actions
Interaction
Agentic
Environment
APIs and tools
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Agents, Vision
Strongest model tested
GPT-5.1 #66 on the capability index
Last model added
November 2025
Scores refreshed
November 20th, 2026
16
Entries on its own leaderboard, 8 models we track

Grok 4 leads with 41.0 among models released by November 2025 Grok 4 leads with 41.0 among models released by November 2025

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. This view shows only models we track, each at its best.

The newest model with a score here came out in November 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Grok 4 41.0
  2. 2 Kimi K2 Sep 2025 30.4
  3. 3 OpenAI deep research 29.3
  4. 4 DeepSeek-V3.1 26.5
  5. 5 GPT-4o mini 23.7
  6. 6 Qwen3-VL 22.9
  7. 7 ASearcher-Web-7B 21.6
  8. 8 Gemini 2.5 pro (snapshot unspecified) 20.5
  9. 9 Gemini 2.5 Flash Sep 2025, thinking 19.3
  10. 10 MMSearch-R1-7B 19.2
  11. 11 WebWatcher 32B 18.7
  12. 12 DR Tulu-8B 18.1
  13. 13 GPT-5.1 17.8
  14. 14 GLM-4.5V 14.4
  15. 15 MiroThinker-v1.0-30B 13.3
  16. 16 Tongyi Deep Research 30B 10.7

Source: Official PaperScope leaderboard

#EntryScoreSettingDateSource
1 Grok 4 41.0 Jul 2025 Official leaderboard
2 Kimi K2 Sep 2025 30.4 ReAct Sep 2025 Official leaderboard
3 OpenAI deep research 29.3 Feb 2025 Official leaderboard
4 DeepSeek-V3.1 26.5 ReAct Aug 2025 Official leaderboard
5 GPT-4o mini 23.7 ReAct Jul 2024 Official leaderboard
6 Qwen3-VL 22.9 ReAct Official leaderboard
7 ASearcher-Web-7B 21.6 Official leaderboard
8 Gemini 2.5 pro (snapshot unspecified) 20.5 ReAct Official leaderboard
9 Gemini 2.5 Flash Sep 2025 19.3 thinking, ReAct Sep 2025 Official leaderboard
10 MMSearch-R1-7B 19.2 Official leaderboard
11 WebWatcher 32B 18.7 ReAct Official leaderboard
12 DR Tulu-8B 18.1 Official leaderboard
13 GPT-5.1 17.8 ReAct Nov 2025 Official leaderboard
14 GLM-4.5V 14.4 ReAct Aug 2025 Official leaderboard
15 MiroThinker-v1.0-30B 13.3 Official leaderboard
16 Tongyi Deep Research 30B 10.7 Official leaderboard
#ModelScoreSettingDateSource
1 Grok 4 41.0 Jul 2025 Official leaderboard
2 Kimi K2 Sep 2025 30.4 ReAct Sep 2025 Official leaderboard
3 OpenAI deep research 29.3 Feb 2025 Official leaderboard
4 DeepSeek-V3.1 26.5 ReAct Aug 2025 Official leaderboard
5 GPT-4o mini 23.7 ReAct Jul 2024 Official leaderboard
6 Gemini 2.5 Flash Sep 2025 19.3 thinking, ReAct Sep 2025 Official leaderboard
7 GPT-5.1 17.8 ReAct Nov 2025 Official leaderboard
8 GLM-4.5V 14.4 ReAct Aug 2025 Official leaderboard

Rankings by part

We read 4 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Reasoning, Induction and Summary. Grok 4 leads 1 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Reasoning

8 models we track have a score here.

  1. 1 Grok 4 36.0
  2. 2 OpenAI deep research ↑1 13.0
  3. 3 Kimi K2 Sep 2025 ↓1 12.0
  4. 4 Gemini 2.5 Flash Sep 2025 ↑2 7.00
  5. 5 GPT-4o mini 6.00
  6. 6 DeepSeek-V3.1 ↓2 6.00
  7. 7 GLM-4.5V ↑1 0.00
  8. 8 GPT-5.1 ↓1 0.00

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

xAI
Own leaderboard Not tested since Jan 2026
Grok 4 41.0

Source: Official PaperScope leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard
10.0 20.0 30.0 40.0 50.0 2025 2026 10.0 20.0 30.0 40.0 50.0 2025 2026

Source: Official PaperScope 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 Tool use

All 265 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 Neon 56.3%
  • MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks. Gemini 4 Argon 90.7%
  • 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%
  • Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
  • DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 12 months ago, in November 2025, for GPT-5.1. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (16 entries).
  • Pass Models scored 8 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 12 months ago, in November 2025, for GPT-5.1.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.1 (#66).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 8 rows for models we track and 8 other entries 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.

How we check