Thinking Machines
Tinker, a post-training API for open-weight models, plus the open Inkling models.
- Founded
- 2025
- Example models
- Inkling, Inkling-Small, Qwen3.5, Kimi K2.6
By The Subconscious Team · Updated
What is Thinking Machines?
Thinking Machines Lab is the San Francisco lab Mira Murati, formerly CTO of OpenAI, founded in February 2025, with OpenAI co-founder John Schulman as chief scientist. It raised about $2 billion at a $12 billion valuation in July 2025 in a round led by Andreessen Horowitz, and in March 2026 signed a multi-year Nvidia deal for one gigawatt of Vera Rubin capacity. Its main developer product is Tinker, an API for post-training open-weight models that launched in October 2025 and is now generally available. Tinker exposes four low-level calls, forward_backward, optim_step, sample and save_state, so teams write their own supervised or reinforcement learning loops while Thinking Machines runs the distributed GPU work. Training uses LoRA adapters rather than full weight updates.
In July 2026 the lab released its own open-weight models under Apache 2.0: Inkling, a 975B-parameter mixture-of-experts model with 41B active, and Inkling-Small at 276B total and 12B active. Both accept text, image and audio input with up to 1M tokens of context. Tinker also trains Qwen3.5, Nemotron 3, GLM-5.3, Kimi K2.6, DeepSeek-V3.1 and gpt-oss. Billing is per million tokens across prefill, sample and train meters; GPT-OSS-20B runs $0.18, $0.45 and $0.40, and cached prefill is 80% off. A beta serverless API serves only the two Inkling models, with Inkling at $1.00 in and $4.05 out. An OpenAI-compatible endpoint can sample any fine-tuned checkpoint, but the docs scope it to testing and low internal traffic.
Thinking Machines pros, cons and use cases
Upsides
- Full control of the training loop, including RL, without managing GPU clusters.
- Fine-tune large MoE models like Kimi K2.6 and Inkling that are hard to train in-house.
- Own Apache 2.0 Inkling models with native image and audio input and 1M context.
- Sample from checkpoints mid-training through an OpenAI-compatible endpoint.
Core use cases
- Research teams running custom SFT or RL post-training on open models.
- Building a task-specialized model on top of an open-weight base.
- Evaluating the Inkling models through the beta serverless API.
Downsides
- Not a general-purpose inference host: serverless covers only Inkling models, and checkpoint sampling is not meant for user-facing traffic.
- LoRA only, and developers write their own training code rather than using a no-code fine-tuning flow.
Thinking Machines alternatives compared
Pick any row for the full head-to-head.
| Provider | Model access | Flagship models | Speed | Price | Customization | Deployment | Long context | Compare |
|---|---|---|---|---|---|---|---|---|
| Thinking Machines | Open weights | Inkling, Inkling-Small | Unknown | Per 1M tokens by prefill, sample, train; Inkling $1.00 in, $4.05 out | LoRA SFT and RL via Tinker | Training API, beta serverless (Inkling only) | Inkling up to 1M; Tinker 32K–256K | |
 Subconscious | Open weights | GLM 5.3, DeepSeek V4.1 Flash | 2x faster task completion | 50–80% lower cost; billed on processed tokens | Marathon post-trained variants | Managed API, dedicated, on-prem | 5M+ effective context | Compare |
 OpenAI | Closed, plus open gpt-oss | GPT-6 Astra, GPT-5.6 Sol, Terra, Luna | Fast mode: up to 2.5x at 2x price | $0.20–$10 in, $1.20–$50 out per 1M | N/A | API, Azure OpenAI, Bedrock | 1.05M; 2x input past 272K | Compare |
 Anthropic | Closed | Claude Fable 5.1, Opus, Sonnet, Haiku 4.5 | Fable is the slowest tier | $1–$10 in, $5–$50 out per 1M | N/A | API, Bedrock, Vertex AI, Microsoft Foundry | 1M, no surcharge past 200K | Compare |
 Google Vertex AI | Closed and open, 200+ models | Gemini 3.8 Flash, Claude, Gemma | Flash tier built for low latency | Gemini 3.8 Flash $0.75 in, $3.75 out | Custom training on GPUs or TPUs | Managed on Google Cloud | 1M on Gemini 3.8 Flash | Compare |
 Amazon Bedrock | Closed and open, 100+ models | Claude, GPT-6 Astra, Nova, DeepSeek | Latency-optimized option on some models | ~20–35% above direct; Claude at parity | Fine-tuning, Custom Model Import | Managed on AWS, AgentCore | Varies by model | Compare |
 Together AI | Open weights | Kimi K3, DeepSeek V4, GLM 5.2, Qwen 3.8 | 0.99s TTFT on DeepSeek V4 Pro | Parity with Fireworks and Baseten | LoRA and full SFT; RL in beta | Serverless, dedicated, GPU clusters | 512K on DeepSeek V4 Pro | Compare |
 Fireworks AI | Open weights | DeepSeek V4 Pro, Kimi K3 | 167–174 tok/s on DeepSeek V4 Pro | Fine-tunes served at base price | SFT, DPO, RFT; Training API | Serverless, dedicated GPUs | Full 1M on DeepSeek V4 Pro | Compare |
 Baseten | Open weights, 13 curated | GLM 5.2, DeepSeek V4, Kimi K3, gpt-oss 120B | 0.49s TTFT, lowest measured | H100 about $6.50/hr dedicated | Deploy any model with Truss | Model APIs, dedicated, self-host | Varies by model | Compare |
 Groq | Open weights | GPT-OSS 120B, Qwen 3.6 27B | 500–1,000 tok/s | Near the floor on small models | No fine-tuned model hosting | GroqCloud API | Around 131K max | Compare |
 Cerebras | Open weights | GPT-OSS 120B, Gemma 4 31B | ~3,000 tok/s on GPT-OSS 120B | $0.35 in, $0.75 out (GPT-OSS 120B) | Unknown | Shared API, dedicated, partners | Unknown | Compare |
 DeepInfra | Open weights | DeepSeek V4 Flash, Llama 3.1 8B | ~33 tok/s on DeepSeek V4 Pro (FP4) | From $0.02 per 1M | No managed fine-tuning | Shared API, no contracts | 66K on FP4 DeepSeek V4 Pro | Compare |
 Hugging Face Inference Providers | Open weights | GLM 5.3, Kimi K3, DeepSeek V4.1 Flash | Routes to fastest provider by default | Provider rates, no markup | N/A | Serverless router; dedicated Endpoints | Up to 1M, provider-dependent | Compare |
 Modal | Bring your own weights | None hosted | ~1s container boot | Per second; H100 $3.95/hr list | Run any training code | Serverless GPU containers | Depends on the model you deploy | Compare |
 Cloudflare Workers AI | Open weights | DeepSeek V4 Pro, GLM 5.3, Kimi K2.7 Code, gpt-oss 120B | Unknown | $0.011 per 1K Neurons; 10K free daily | BYO LoRA on small models (beta) | Serverless on Cloudflare network | 1M on DeepSeek V4; 262K on Kimi | Compare |
 xAI | Closed | Grok 4.6, Grok 4.20, grok-build | ~54 tok/s on Grok 4.6 | $2 in, $6 out (Grok 4.6); 2x past 200K | Unknown | First-party API | 500K (4.6), 1M (4.20, 4.3) | Compare |
 Mistral AI | Open weights, plus closed Codestral | Mistral Medium 3.5, Small 4, Large 3 | Unknown | $0.15–$1.50 in, $0.60–$7.50 out per 1M | Forge (enterprise); fine-tuning API deprecated | API, Azure, Bedrock, Vertex, self-host | 256K | Compare |
 DeepSeek | Open weights (MIT) | DeepSeek V4.1 Flash, V4 Pro | ~35 tok/s on V4 Pro | Off-peak hours at half price | Open weights to fine-tune | First-party API, Hugging Face weights | 1M, 384K max output | Compare |
 Moonshot AI | Open weights, custom license | Kimi K3, Kimi K2.6 | ~33 tok/s on Kimi K3 | $3 in, $15 out (Kimi K3) | Open weights to fine-tune | API, Kimi Code, OpenRouter | 1M | Compare |
 Z.ai | Open weights (MIT) | GLM-5.3, GLM-5.3-Flash | ~80 tok/s on GLM-5.3 | $1.40 in, $4.40 out (GLM-5.3); free Flash tier | Open weights, no license limits | API, GLM Coding Plan | 1M (GLM-5.3) | Compare |
 Alibaba Cloud | Closed Max; open smaller Qwen | Qwen 3.8-Max, Qwen 3.7-Max | ~40 tok/s on Qwen 3.8-Max | $2 in, $6 out international | No fine-tuning on Max | Model Studio on Alibaba Cloud | 1M (Qwen 3.8-Max) | Compare |
 Meta | Closed API; open Muse Glimmer | Muse Spark 1.3, Muse Glimmer | ~145–233 tok/s on Muse Spark 1.3 | $1.25 in, $4.25 out; Contributor tier cheaper | Open Muse Glimmer weights to fine-tune | Meta Model API (preview) | 1M | Compare |
 Cohere | Closed, plus open Command A+ | Command A+, Command A, Embed 4, Rerank 4 | 375 tok/s on Command A+ W4A4, per Cohere | $0.0375–$2.50 in, $0.15–$10 out per 1M | Enterprise fine-tuning, incl. private | API, Bedrock, Azure, OCI, VPC, on-prem | 256K on Command A; 128K on A+ | Compare |
 SambaNova | Open weights | MiniMax M2.7, GPT-OSS 120B, DeepSeek | ~820 tok/s on MiniMax M2.7 (SN50) | $0.22 in, $0.59 out (GPT-OSS 120B) | Unknown | SambaCloud, racks for neoclouds | Up to 192K (MiniMax M2.7) | Compare |
 Nebius | Open weights, 60+ models | DeepSeek, Qwen, GLM, Kimi, GPT-OSS | Among top hosts on throughput | From $0.06 per 1M input | Serve uploaded fine-tunes | Token Factory, dedicated, raw GPUs | Varies by model | Compare |
 Crusoe | Open weights | DeepSeek V4, GLM 5.3, Kimi K2.6, Nemotron 3 | Up to 9.9x faster TTFT vs vLLM (vendor claim) | $0.05–$1.74 in, $0.20–$4.40 out per 1M | Serverless LoRA fine-tuning | Serverless, self-serve and tailored dedicated, raw GPUs | Varies by model; cluster-wide KV cache | Compare |
 fal | Hosted media models | FLUX, Kling, Seedream | Cold starts on less popular endpoints | Per image, per video second, GPU time | LoRA training endpoints | Hosted API, serverless GPUs | Not applicable | Compare |
 Novita AI | Open weights | DeepSeek V4 Pro, Gemma 4 | ~36 tok/s on DeepSeek V4 Pro | From $0.02 per 1M; batch 50% off | Hot-swappable LoRA adapters | Serverless, GPU cloud, dedicated | Full 1M on DeepSeek V4 Pro | Compare |
 Venice | Open weights, plus proxied closed models | GLM 5.3, Kimi K3, DeepSeek V4 Pro | Unknown | $0.06–$12 in, $0.28–$60 out per 1M; DIEM staking | Unknown | Serverless API, consumer app | 1M on most current models | Compare |
 Parasail | Any Hugging Face model | GTE-Qwen2, Qwen3-VL-8B-Instruct | 600ms p99 real-time budget | Per-parameter rates; batch 50% off | Private Hugging Face repos | Serverless, elastic, dedicated, batch | Varies by model | Compare |
 Inference.net | Open, closed and custom | Customer fine-tunes | Batch windows of 24h to 7 days | Discounted spare GPU capacity | Distill traces into custom models | Batch API, gateway, dedicated GPUs | Varies by model | Compare |
 GMI Cloud | Open and third-party models | GLM-4.7-Flash, Google Veo | Near bare-metal performance | $0.07 in, $0.40 out (GLM-4.7-Flash) | Unknown | Shared, autoscaling, reserved GPUs | Varies by model | Compare |
 Sail Research | Open weights | Kimi K2.6, GLM-5, GPT-OSS 120B | Minutes per turn by design | 30–80% off by completion window | Customer LoRA fine-tunes | API plus Sailboxes | Varies by model | Compare |
 Morph | Specialist models | morph-v3-fast, morph-v3-large | 10,500+ tok/s Fast Apply | ~40% fewer tokens than full rewrites | Fine-tuning offered | OpenAI-compatible API | Unknown | Compare |
 Relace | Specialist models | relace-apply-3, agentic search | ~10,000 tok/s apply | 3x+ cheaper than full rewrites | Unknown | Hosted API or self-hosted | 128K max | Compare |
 TypeSafe AI | Decision models | Jev, jev-1.13 | ~100ms per call | A fraction of an LLM call | Unknown | Early-access API | Unknown | Compare |
 StepFun | Open (Apache 2.0) and API models | Step 3.7 Flash, Step3 | ~128 tok/s on Step 3.7 Flash | $0.20 in, $1.15 out (Step 3.7 Flash) | Open weights to fine-tune | First-party API, OpenRouter | 256K | Compare |
 Runware | Hosted media models | Seedance 2.5, Qwen-Image-3.0 | Unknown | Images from fractions of a cent | Fine-tuned diffusion checkpoints | Unified API, raw GPUs | Not applicable | Compare |
 StreamLake | Proprietary coding models | KAT-Coder-Pro V2.5, KAT-Coder-Air | Unknown | Per token or KwaiKAT Coding Plan | Unknown | MaaS API, bare metal | Unknown | Compare |
 Wafer | Open weights | Qwen 3.5 397B Turbo, GLM 5.1 Turbo | 2–2.8x vs stock vLLM or SGLang | Wafer Pass from $10 a week | Agent-tuned dedicated deployments | Serverless pass, dedicated | Varies by model | Compare |
 RunInfra | Open weights | Nemotron 3.5 Lightning 30B, Qwen 3.8 27B | Cold starts under 2s | Coding plans from $10 a month | Uploads up to 50 GB; auto-quantization | Model APIs, agent-built endpoints | Varies by model | Compare |
 Particle.AI | Open weights | DeepSeek V4.1 Flash, GLM 5.3 Flash | ~157 tok/s on DeepSeek V4.1 Flash | $0.10 in, $0.40 out (GLM 5.3 Flash) | Unknown | Via Vercel AI Gateway | 1M | Compare |
Frequently asked questions
What is Thinking Machines?
Thinking Machines Lab is the San Francisco lab Mira Murati, formerly CTO of OpenAI, founded in February 2025, with OpenAI co-founder John Schulman as chief scientist. It raised about $2 billion at a $12 billion valuation in July 2025 in a round led by Andreessen Horowitz, and in March 2026 signed a multi-year Nvidia deal for one gigawatt of Vera Rubin capacity. Its main developer product is Tinker, an API for post-training open-weight models that launched in October 2025 and is now generally available. Tinker exposes four low-level calls, forward_backward, optim_step, sample and save_state, so teams write their own supervised or reinforcement learning loops while Thinking Machines runs the distributed GPU work. Training uses LoRA adapters rather than full weight updates.
What is Thinking Machines best for?
Research teams running custom SFT or RL post-training on open models; Building a task-specialized model on top of an open-weight base; Evaluating the Inkling models through the beta serverless API.
How much does Thinking Machines cost?
Thinking Machines pricing at a glance: Per 1M tokens by prefill, sample, train; Inkling $1.00 in, $4.05 out. Rates change often, so check Thinking Machines's pricing page before committing.
How much context does Thinking Machines support?
Thinking Machines's long-context support: Inkling up to 1M; Tinker 32K–256K.
What are the downsides of Thinking Machines?
Not a general-purpose inference host: serverless covers only Inkling models, and checkpoint sampling is not meant for user-facing traffic; LoRA only, and developers write their own training code rather than using a no-code fine-tuning flow.
What are the best alternatives to Thinking Machines?
Common alternatives include Subconscious, OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock. Each has a head-to-head comparison with Thinking Machines on this site.
Sources: Tinker, Tinker models and pricing, Inkling model card, Tinker OpenAI-compatible API. Pricing and model lineups change often; figures are a snapshot.