Reka Core / Flash / Edge
SF / Singapore foundation-model startup founded by ex-DeepMind / Meta / Google researchers (Dani Yogatama, Yi Tay, Qi Liu). Native multimodal models — Reka Core (~67B, frontier candidate), Flash (~21B), Edge (~7B). Apr-2024: claimed Core matched GPT-4 / Gemini Ultra on MMMU. Acquired by Snowflake announced May-2025 (~$2B reported). Distinct from frontier-tier in mindshare but important as multimodal-native open-weights option.
At a glance
- Type
- Multimodal foundation model family — Reka Core / Flash / Edge
- Tier
- T2
- Created
- 2022-07 (Reka founded); 2024-04 (Reka Core launched)
- Latest release
- Reka Flash 3 (2025); Reka Core (2024-04)
- License
- Reka Flash 3 weights under Apache 2.0; Core proprietary
- Pricing
- Reka API pay-per-token (~$3/$10 per 1M Core); post-Snowflake-acq: bundled into Snowflake Cortex AI
- Funding
- $103M total raised (Series B June-2023 led by DST Global + Snowflake Ventures); Snowflake acquisition reported May-2025 (~$2B)
Taxonomy
- storage
- parametric
- retrieval
- parametric-recall
- persistence
- parametric-permanent
- update
- read-only
- unit
- weight
- governance
- opaque
- conflict
- n/a
When to use
Optimised for: multimodal-native (single-model image+text+video understanding); APAC / Snowflake-customer enterprise
Anti-fit: not for users wanting frontier-tier raw benchmark — now lags Claude/GPT/Gemini; Snowflake acquisition uncertainty re. standalone API future
Pros & cons
Pros
Native multimodal training (not late-fusion); strong DeepMind / Google / Meta founding team; Snowflake distribution post-acquisition; Singapore HQ for APAC compliance.
Cons
Lags frontier tier on raw benchmarks post-2024; Snowflake-acquisition integration uncertainty; smaller ecosystem; limited model releases since 2024.
Claims & capabilities
Reka Core matched GPT-4 / Gemini Ultra on MMMU (Apr-2024 vendor claim); native multimodal training (no late-fusion); 128k context; Snowflake acquisition announced May-2025 (~$2B reported)
Technical surface
- API surface
- REST + Python SDK; OpenAI-compatible
- Backend storage
- not applicable — substrate foundation model
- Deployment
- Managed cloud (Reka API) + Snowflake Cortex AI post-acquisition; some weights via HuggingFace (Reka Flash 3)
- Embedding model
- not applicable — not a memory product
- Multi-tenancy
- not applicable — substrate foundation model
- MCP
- not applicable — substrate foundation model
- A2A
- not applicable — substrate foundation model
- OpenTelemetry
- not applicable — substrate foundation model
Similar systems
Other foundation models (substrate reference) in the catalog, ranked by inbound references.
- Anthropic Claude (foundation models) T1
Claude family of frontier foundation models — Sonnet 4 (workhorse), Opus 4.5 (frontier capability), Haiku 4.5 (fast/cheap). Frontier model used as substrate by Claude Code, Anthropic Memory tool, Cline, Aider, Continue.dev, Cursor, Goose, and dozens of other catalog entries. Memory in Claude is parametric (in weights) — discrete episodic memory lives outside the model in the Claude Memory tool (see Platform-provider memory section).
- OpenAI GPT family (GPT-5 / GPT-4o / o3 / o4) T1
OpenAI's foundation model family — GPT-5 (2025-08 flagship), GPT-4o (multimodal workhorse), o3/o4 reasoning series. Substrate for ChatGPT, the OpenAI API, Microsoft Copilot, Azure OpenAI, and many downstream agents in this catalog (Devin, Codex CLI, Operator, Cursor mode). Parametric memory in weights; discrete user memory ships in ChatGPT Memory feature (separate row).
- Google Gemini 3 family T1
Google DeepMind's Gemini 3 frontier model family — Gemini 3 Pro (top tier), 3 Flash (fast), 3 Nano (on-device). Substrate for Google Workspace AI, Vertex AI Agent Platform, Gemini Code Assist, Project Mariner, and Project Astra. Multimodal-native; 1M-10M context window. Largest context window of any frontier model. Used by Gemini Enterprise Agent Platform Memory Bank (separate row).
- Mistral Large 2 / Mixtral family T1
French frontier-model lab (Paris). Family includes Mistral Large 2 (123B dense, Jul-2024), Mixtral 8x22B (open-weights MoE), Mistral Small 3 (24B, open-weights Jan-2025), Codestral (code specialist). EU-headquartered alternative to US/China model labs; GDPR-friendly data residency. Distinct from the Mistral Le Chat memory feature (already in catalog) — this row covers the model family itself as substrate.
- Meta Llama 4 family T1
Meta's Llama 4 family (released 2025-04) — Scout (109B MoE, 17B active), Maverick (400B MoE, 17B active), Behemoth (2T MoE, 288B active; preview). First Llama generation to use mixture-of-experts and to be multimodal-native. Substrate for many OSS agents (Open-Interpreter, Ollama-served local agents, vLLM-served enterprise inference). Llama community license (commercial use permitted under 700M MAU threshold).
- xAI Grok 4 T1
Elon Musk's xAI frontier model — Grok 4 launched 2025-07. Trained on Colossus (100k+ GPU supercluster in Memphis). Deeply integrated with X/Twitter (real-time tweet data access). Grok 4 Heavy variant for premium tier. No frontier-tier open-weights — Grok 1 (314B MoE) is the only weight release.