RL environments and AI benchmarks, cataloged daily.

Who builds RL environments, who funds them and how models score, rebuilt every morning from public sources. Every placement links to the line on the company's own site that puts it there.

Today in RL

The headline addition is a full suite of nine Jais-3 models from IFM, spanning 0.9B to 375B parameters and including a mixture-of-experts vision variant and a dedicated speech-recognition model. Meta added Muse Spark Open (62.66), and NVIDIA expanded its Nemotron 3.5 Super VL line with two 120B vision-language variants. Liquid AI quietly entered a new model, d1-ultra, without a published score. On the leaderboard, scores moved broadly upward across the top 30, with Claude Opus 5.5 edging up to 86.49, GPT-6.1 Sol climbing to 84.04, and Gemini 4 Argon gaining the most ground in the top ten, reaching 80.96. Nine new benchmarks also joined the field.

  • IFM Launches Jais-3 Family

    Nine Models Spanning 0.9B to 375B: IFM debuted the Jais-3 series today — nine models covering dense sizes at 0.9B, 3.7B, 7B, and 32B, each paired with an 'Uno' variant, plus a 375B mixture-of-experts model (Jais-3-375B-A23B with 23B active parameters), a 36B MoE vision model (Jais-3-MoVA-36B-A4B), and a dedicated automatic speech recognition model (Jais-ASR). The breadth of the release — spanning edge, mid-tier, frontier, multimodal, and speech — signals IFM is positioning Jais-3 as a full-stack model family.

  • NVIDIA

    NVIDIA Adds Two 120B Vision-Language Variants to Nemotron 3.5 Super VL

    NVIDIA expanded its Nemotron 3.5 Super VL line with two new 120B 12B-active-parameter models: a standard version scoring 49.51 and an NVFP4 quantized variant. The NVFP4 release points to growing emphasis on deployment-optimized frontier VL models that can run at lower precision without a full separate training run.

  • Meta

    Meta Adds Muse Spark Open at 62.66

    Meta introduced Muse Spark Open, an open-weight entry in the Muse Spark line, scoring 62.66. This places it below the existing Muse Spark 1.1 (65.99) and Muse Spark 1.2 (65.85), suggesting it is a lighter or less capable open release relative to the closed Muse Spark variants already on the leaderboard.

  • Liquid AI Enters d1-ultra Without a Published Score

    Liquid AI added d1-ultra to the leaderboard today without an overall score, continuing a pattern seen with other labs releasing models ahead of full benchmark evaluations. Liquid AI previously debuted LFM2.5-24B-A2B on October 14.

  • Broad Score Lift Across the Top 30 — Gemini 4 Argon Gains Most

    Today's score refresh pushed most top-30 models upward by small but consistent margins. Gemini 4 Argon gained the most at the top, rising from 80.87 to 80.96. GPT-6.1 Sol climbed from 83.95 to 84.04, and GPT-6 Astra moved from 83.75 to 83.86. The few models that dipped — Claude Haiku 5.5 (70.16 → 70.07), Claude Opus 4.7 (66.86 → 66.83), GPT-5.5 (69.38 → 69.35) — fell by less than 0.1 points, making this one of the more uniformly positive refresh days in recent weeks.

The catalog

The top of each list. Every list has a full table, sortable and filterable.

Models

By overall score

  1. 1Claude Opus 5.586.49
  2. 2GPT-6.1 Sol84.04
  3. 3GPT-6 Astra83.86
  4. 4Claude Fable 5.183.18
  5. 5Claude Sonnet 5.582.14

All 49 models

RL environment companies

By tier, X reputation, benchmarks built

  1. SSurge AI
  2. SSnorkel AI
  3. SPrime Intellect
  4. SScale AI
  5. SMercor

All 70 companies

Benchmarks

S tier: the ones frontier labs report

  1. SHumanity's Last Exam
  2. STerminal-Bench 2.1
  3. SSWE-bench Verified
  4. SAIME 2025
  5. SBrowseComp

All 3,089 benchmarks

Infrastructure companies

By tier, X reputation, benchmarks built

  1. SPrime Intellect
  2. SModal
  3. AE2B
  4. ADaytona
  5. BVeris AI

All 11 companies

Browse by domain

All 91 domains and labels

On the map. Our domains: every company and benchmark sits in one, placed by a quote from its own site or paper.

Directory labels. Tags other catalogs put on entries. Not domains on the map.

The other 67 pages

More pages for domains on the map

More directory labels

The map

Every company we track, placed in the domain it leads with. Click one for the line on its own website that puts it there.

All companies

Every company we track, whatever it sells, placed in the domain it leads with. Companies that cover many domains with none leading sit in the middle.

77companies
15domains
1with no leader

Domains

Multi-domain30cover many domains Coding1313 lead · 32 also cover Infrastructure6runtime, no domain Enterprise work55 lead · 29 also cover Security44 lead · 10 also cover Computer use33 lead · 13 also cover Finance33 lead · 21 also cover Health and bio33 lead · 14 also cover Media33 lead · 9 also cover Robotics22 lead · 9 also cover Science22 lead · 17 also cover Chip design11 lead · 1 also cover Games11 lead · 5 also cover Math11 lead · 9 also cover Legal00 lead · 12 also cover Not yet confirmed5listed, not yet confirmed
Every vendor, with quotes →
Glow · what it leads with selling

The galaxy could not load. See every vendor on the board.

How to read it

A company, in the domain its own site leads with. The logo is its own; the glow is what it mainly sells.

Also covers: a ring at a domain's edge for a company that leads with another domain but also covers this one. A faint line runs to where it leads.

Dots around it: the other things it sells. Hollow while it is still expanding into them.

Click any star for the lines from its own website that put it there.

The map on its own page, with every company listed →

Research

All research posts
  1. Five people, about $300k, synthetic RL environments: Ultramarine 1 Flash beats Claude Fable 5 on 4 of 6 knowledge-work benchmarks
  2. Why expert-data vendors are becoming research labs: 14 studies from four vendors in 40 days
  3. Mercor hired the first authors of LoRA, τ²-bench, SciCode, and APEX-Agents. micro1 is offering up to $4M for a robotics research director
  4. 2,209 agents rewrote a coding harness in Rust in two weeks. The reward was a parity check in a fresh sandbox

In short

RL environments, in short

What are RL environments?
A simulated workplace where an AI agent practises a task (a code repository, a browser, business software) and an automatic grader scores each attempt. Labs train on thousands of these attempts. Bought to improve models. More in the glossary.
Who sells RL environments?
70 companies we track sell RL environments, evals or verifiers, among them AfterQuery, HUD, Mercor, Symbal, Turing, Align AI, BenchFlow and Cua. Each is placed by a line from its own website; see them all on the RL environments page.
What is the difference between an RL environment and a benchmark?
An RL environment is for training: a model practises in it and a grader scores each attempt. A benchmark or eval is for measuring: a fixed, held-out set of tasks with fixed scoring, used to measure and compare models. Never trained on, so the score means something. Used to measure models.
What are the main AI benchmarks for agents and LLMs?
Among the most referenced: Humanity's Last Exam, Terminal-Bench 2.1, SWE-bench Verified, AIME 2025, BrowseComp, Terminal-Bench Hard, GPQA Diamond and GPQA-D. See the full AI benchmarks list.
Where does this data come from?
Every placement is read from the company's own website or the benchmark's own page or paper, quoted and linked, and the catalog is rebuilt daily from public sources. How the data is built.

From the builders

What vendors are posting

The newest post on X from each RL environment vendor we track, refreshed with every build. Scroll the feed for more. How we find their accounts.

  1. Abundant @abundant_labs

  2. Prime Intellect @PrimeIntellect

  3. ReasonCore @reasoncoreai

  4. Vals AI @ValsAI

  5. Modal @modal

  6. micro1 @micro1_ai

  7. Snorkel AI @snorkelai

  8. Cua @trycua

  9. Mercor @mercor

  10. AfterQuery @afterquery