DeepSeek
Anthropic
xAI
Google
Meta
OpenAI
MistralResearch retrieval or analysis?
Search appears as an explicit task for only one model on this shortlist. For every other model, retrieval tooling and citation behavior require separate verification.
Purpose guide · Reasoning and research
Which AI models belong on a reasoning or research shortlist?
Catalog current through July 21, 2026
61 models
Five source-backed models whose Epoch records explicitly name reasoning, proof, or search work. This is a shortlist, not a performance ranking.
Selection rule
Included when the Epoch record explicitly lists mathematical reasoning, quantitative reasoning, theorem proving, proof generation, proof verification, or search. A matching label establishes task relevance only, not quality.
Task-fit shortlist
Evidence first. Your evaluation next.
Ordered by release date for readability, newest first. The order is not a score or recommendation.
- 01
- Matching evidence
- Mathematical reasoning · Theorem proving · Proof generation · Proof verification
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 02
- Matching evidence
- Quantitative reasoning · Search · Mathematical reasoning
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 03
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 04
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 05Moonshot AI
Kimi K2 Thinking
Released November 6, 2025
- Matching evidence
- Search · Quantitative reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 06Alibaba
Tongyi DeepResearch
Released October 28, 2025
- Matching evidence
- Quantitative reasoning · Search
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 07MiniMax
MiniMax-M2
Released October 27, 2025
- Matching evidence
- Search
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 08
- Matching evidence
- Quantitative reasoning · Mathematical reasoning
- Domain
- Language · Vision · Multimodal
- Access
- API access; closed weights
- 09
- Matching evidence
- Search
- Domain
- Vision · Language · Speech
- Access
- API access; closed weights
- 10Alibaba
AgentFounder-30B
Released September 16, 2025
- Matching evidence
- Quantitative reasoning · Mathematical reasoning · Search
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 11Alibaba
Qwen3-Max
Released September 5, 2025
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 12
- Matching evidence
- Search · Quantitative reasoning · Mathematical reasoning
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 13
- Matching evidence
- Mathematical reasoning
- Domain
- Language · Multimodal · Vision · Video · Audio · Mathematics
- Access
- Hosted access (no API); closed weights
- 14Alibaba
Qwen3-235B-A22B (Jul 2025)
Released July 25, 2025
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 15Alibaba
Qwen3-235B-A22B-Thinking (Jul 2025)
Released July 25, 2025
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 16Moonshot AI
Kimi K2
Released July 11, 2025
- Matching evidence
- Quantitative reasoning · Search
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 17
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- Unreleased; closed weights
- 18
- Matching evidence
- Search · Quantitative reasoning
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 19
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Vision · Video · Multimodal · Speech
- Access
- API access; closed weights
- 20
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 21
- Matching evidence
- Quantitative reasoning · Search · Mathematical reasoning
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 22
- Matching evidence
- Quantitative reasoning · Search · Mathematical reasoning
- Domain
- Language · Multimodal · Vision
- Access
- API access; closed weights
- 23
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Vision · Video · Multimodal · Speech
- Access
- API access; closed weights
- 24Alibaba
Qwen3-235B-A22B
Released April 29, 2025
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 25
- Matching evidence
- Quantitative reasoning
- Domain
- Multimodal · Language · Vision
- Access
- Unreleased; closed weights
- 26
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Vision · Video · Multimodal · Speech
- Access
- API access; closed weights
- 27
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 28
- Matching evidence
- Quantitative reasoning · Mathematical reasoning
- Domain
- Language · Mathematics · Multimodal
- Access
- API access; closed weights
- 29Baidu
ERNIE-4.5-VL-424B-A47B (文心大模型4.5)
Released March 16, 2025
- Matching evidence
- Quantitative reasoning
- Domain
- Multimodal · Language · Vision · Video
- Access
- Open weights (unrestricted); open weights
- 30Tencent
Hunyuan-TurboS
Released March 11, 2025
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 31Alibaba
QwQ-32B
Released March 6, 2025
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 32
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Vision · Multimodal
- Access
- API access; closed weights
- 33
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Vision · Multimodal
- Access
- API access; closed weights
- 34
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 35Moonshot AI
Kimi k1.5
Released January 22, 2025
- Matching evidence
- Quantitative reasoning
- Domain
- Multimodal · Language · Vision
- Access
- API access; closed weights
- 36
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 37
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 38
- Matching evidence
- Quantitative reasoning · Search
- Domain
- Language · Vision · Multimodal
- Access
- API access; closed weights
- 39
- Matching evidence
- Quantitative reasoning · Mathematical reasoning
- Domain
- Language · Mathematics · Multimodal
- Access
- API access; closed weights
- 40
- Matching evidence
- Mathematical reasoning
- Domain
- Multimodal · Language · Vision
- Access
- Open weights (non-commercial); open weights
- 41Moonshot AI
k0-math
Released November 16, 2024
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Mathematics · Language
- Access
- Unreleased; closed weights
- 42NVIDIA
NVLM-D 72B
Released October 22, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Vision · Language
- Access
- Open weights (non-commercial); open weights
- 43NVIDIA
NVLM-H 72B
Released October 22, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Vision · Language
- Access
- Unreleased; closed weights
- 44NVIDIA
NVLM-X 72B
Released October 22, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Vision · Language
- Access
- Unreleased; closed weights
- 45Alibaba
Qwen2.5 Instruct (72B)
Released September 19, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 46Alibaba
Qwen2.5-72B
Released September 19, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 47Alibaba
Qwen2.5-32B
Released September 17, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 48
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 49
- Matching evidence
- Quantitative reasoning
- Domain
- Language · Mathematics · Biology
- Access
- API access; closed weights
- 50
- Matching evidence
- Mathematical reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 51
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Mathematics · Language
- Access
- Open weights (unrestricted); open weights
- 52Z.ai
GLM-4 (0520)
Released May 20, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- API access; closed weights
- 53Alibaba
Qwen1.5-72B
Released February 4, 2024
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (restricted use); open weights
- 54
- Matching evidence
- Quantitative reasoning
- Domain
- Language
- Access
- Open weights (unrestricted); open weights
- 55
- Matching evidence
- Mathematical reasoning
- Domain
- Language · Biology
- Access
- Open weights (non-commercial); open weights
- 56
- Matching evidence
- Mathematical reasoning
- Domain
- Language
- Access
- Unreleased; closed weights
- 57
- Matching evidence
- Mathematical reasoning · Quantitative reasoning
- Domain
- Language
- Access
- Unreleased; closed weights
- 58
- Matching evidence
- Quantitative reasoning · Mathematical reasoning
- Domain
- Language
- Access
- Unreleased; closed weights
- 59
- Matching evidence
- Mathematical reasoning
- Domain
- Language · Mathematics
- Access
- API access; closed weights
- 60Microsoft
MEB
Released September 4, 2021
- Matching evidence
- Search
- Domain
- Search · Language
- Access
- Hosted access (no API); closed weights
- 61Microsoft
RankNet
Released August 7, 2005
- Matching evidence
- Search
- Domain
- Search
- Access
- Unreleased; closed weights
Decision criteria
Questions the source can help you ask.
These checks narrow an evaluation plan. They do not replace hands-on testing.
General reasoning or formal proof?
Use the task labels to distinguish quantitative reasoning from theorem and proof work. Do not assume one label covers the other.
What must you test yourself?
Evaluate factual accuracy, source fidelity, calibration, tool use, latency, context handling, and cost on your own workload. This catalog does not measure them.
Evidence boundary
What this guide does not know.
No task-level quality, price, context window, latency, or reliability measurement is present in the Epoch catalog. We therefore publish no ranking or “best” label.
Catalog source: Epoch AI Data on AI Models, retrieved 2026-07-21, CC-BY. Artifact sha256:80ad08198183693101ceb246a5ea272d64c06c147fc41f72cca8526b32483424.
Each listed model links to its model-level source above. See the publication source ledger and methodology for the wider evidence policy.