What's Changed

• launch: skip unchanged integration rewrite configration by @hoyyeva in #15491
• Revert "launch/opencode: use inline config" by @hoyyeva in #15568
• launch/openclaw: fix --yes flag behaviour to skip channels configuration by @hoyyeva in #15589
• launch: OpenCode inline config by @hoyyeva in #15586
• Add MLX closure support by @jessegross in #15590
• mlx: Improve gemma4 performance with fused operations by @dhiltgen in #15587
• mlx: fix RotatingKVCache.concat() dropping context on mid-rotation by @dhiltgen in #15591
• launch: add hermes by @ParthSareen in #15569
• launch: always list cloud recommendations first by @hoyyeva in #15593
• Keep Gemma4 router projection in source precision by @dhiltgen in #15613
• gemma4: render differently based on model size by @drifkin in #15612
• cmd/launch: add Copilot CLI integration by @scaryrawr in #15583
• mlx: fix imagegen lookup by @dhiltgen in #15588
• mlx: fix gemma4 cache to use logical view by @dhiltgen in #15617

New Contributors

• @scaryrawr made their first contribution in #15583

Full Changelog: v0.20.8-rc0...v0.21.0-rc0

What's Changed

• ROCm: Update to ROCm 7.2.1 on Linux by @saman-amd in #15483
• gemma4: fix nothink case renderer by @drifkin in #15553
• gemma4: fix compiler error on metal by @dhiltgen in #15550
• gemma4: add nothink renderer tests by @drifkin in #15554
• mlx: mixed-precision quant and capability detection improvements by @dhiltgen in #15409
• mlx: add op wrappers for Conv2d, Pad, activations, trig, and masked SDPA by @dhiltgen in #14913
• Revert "gemma4: add nothink renderer tests" by @drifkin in #15555
• cgo: suppress deprecated warning to quiet down go build by @dhiltgen in #15438
• mac: prevent generate on cross-compiles by @dhiltgen in #15120
• Revert "gemma4: fix nothink case renderer" by @drifkin in #15556
• launch/opencode: use inline config by @hoyyeva in #15462
• gemma4: restore e2b-style nothink prompt by @drifkin in #15560
• Gemma4 on MLX by @dhiltgen in #15244

Full Changelog: v0.20.6...v0.20.8-rc0

What's Changed

• Fix quality of gemma:e2b and gemma:e4b when thinking is disabled
• ROCm: Update to ROCm 7.2.1 on Linux by @saman-amd in #15483

Full Changelog: v0.20.6...v0.20.7

What's Changed

• Gemma 4 tool calling ability is improved and updated to use Google's latest post-launch fixes
• Parallel tool calling improved for streaming responses
• Hermes agent Ollama integration guide is now available
• Ollama app is updated to fix image attachment errors

New Contributors

@matteocelani made their first contribution in #15272

Full Changelog: v0.20.5...v0.20.6

OpenClaw channel setup with ollama launch

What's Changed

• OpenClaw channel setup: connect WhatsApp, Telegram, Discord, and other messaging channels through ollama launch openclaw
• Enable flash attention for Gemma 4 on compatible GPUs
• ollama launch opencode now detects curl-based OpenCode installs at ~/.opencode/bin
• Fix /save command for models imported from safetensors

New Contributors

• @sjhddh made their first contribution in #15424

Full Changelog: v0.20.4...v0.20.5

What's Changed

• mlx: Improve M5 performance with NAX
• gemma4: enable flash attention

Full Changelog: v0.20.3...v0.20.4

What's Changed

• Gemma 4 Tool Calling improvements
• Added latest models to Ollama App
• OpenClaw fixes for launching TUI

Full Changelog: v0.20.2...v0.20.3

What's Changed

• app: default app home view to new chat instead of launch by @jmorganca in #15312

Full Changelog: v0.20.1...v0.20.2

What's Changed

• bench: add prompt calibration, context size flag, and NumCtx reporting by @dhiltgen in #15158
• model/parsers: fix gemma4 arg parsing when quoted strings contain " by @drifkin in #15254
• ggml: skip cublasGemmBatchedEx during graph reservation by @jessegross in #15301
• gemma4: enable flash attention by @dhiltgen in #15296
• ggml: fix ROCm build for cublasGemmBatchedEx reserve wrapper by @jessegross in #15305
• model/parsers: rework gemma4 tool call handling by @drifkin in #15306

Full Changelog: v0.20.0...v0.20.1

Gemma 4

Effective 2B (E2B)

ollama run gemma4:e2b

Effective 4B (E4B)

ollama run gemma4:e4b

26B (Mixture of Experts model with 4B active parameters)

ollama run gemma4:26b

31B (Dense)

ollama run gemma4:31b

What's Changed

• docs: update pi docs by @ParthSareen in #15152
• mlx: respect tokenizer add_bos_token setting in pipeline by @dhiltgen in #15185
• tokenizer: add SentencePiece-style BPE support by @dhiltgen in #15162

Full Changelog: v0.19.0...v0.20.0-rc0