# Raw Qwen3 4B Instruct 2507 direct logits vs typecastlm (Mikhail Gribov, Qwen3.5-4B computed head)

> Side-by-side benchmark comparison for Raw Qwen3 4B Instruct 2507 direct logits and typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/jevbench-raw-qwen3-4b-instruct-2507-vs-jevbench-typecastlm
- Last updated: September 30, 2026

- Shared sourced benchmarks: 2
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

Evidence is too thin for a benchmark verdict: Raw Qwen3 4B Instruct 2507 direct logits and typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) share only 2 sourced benchmark rows.

## Summary

- Only 2 shared sourced benchmark rows are available. That is not enough for a broad capability verdict, so this mirror does not name an overall winner.

## Model Snapshot

| Property | Raw Qwen3 4B Instruct 2507 direct logits | typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) |
|----------|----------|----------|
| Creator | Alibaba Qwen / neutral reproduction | Mikhail Gribov |
| Type | Open Weight | Open Weight |
| Reasoning | Non-Reasoning | Non-Reasoning |
| Context | null | null |
| Independent public score (not head-to-head) | null | null |
| Benchmarks Covered | 2 | 2 |
| Pricing (input/output) | Pricing unavailable / Pricing unavailable | Pricing unavailable / Pricing unavailable |

## Category Breakdown

### Reasoning

- Winner: Insufficient shared evidence
- Raw Qwen3 4B Instruct 2507 direct logits public-lane score: Coming soon
- typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) public-lane score: Coming soon

| Benchmark | Raw Qwen3 4B Instruct 2507 direct logits | typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) | Winner |
|-----------|-----------|-----------|--------|
| JevBench 1.4 | 40.95 | 25.28 | Raw Qwen3 4B Instruct 2507 direct logits |
| JevBench 1.5 | 62.14 | 21.83 | Raw Qwen3 4B Instruct 2507 direct logits |

## FAQ

### Is there enough evidence to pick between Raw Qwen3 4B Instruct 2507 direct logits and typecastlm (Mikhail Gribov, Qwen3.5-4B computed head)?

Not yet. The models share only 2 sourced benchmark rows, so this mirror reports the available measurements and metadata without naming a broad winner.

### What can I compare safely on this page?

Published pricing, context windows, model type, and individual shared benchmark rows are directly comparable. Independent overall scores and non-overlapping benchmark coverage are labeled separately and should not be read as a controlled head-to-head result.

## Related Comparisons

- [Raw Qwen3 4B Instruct 2507 direct logits vs GPT-6 Astra](/compare/gpt-6-astra-vs-jevbench-raw-qwen3-4b-instruct-2507)
- [typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) vs GPT-6 Astra](/compare/gpt-6-astra-vs-jevbench-typecastlm)
- [Raw Qwen3 4B Instruct 2507 direct logits vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-jevbench-raw-qwen3-4b-instruct-2507)
- [typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) vs Claude Opus 5.5](/compare/claude-opus-5-5-vs-jevbench-typecastlm)

## Explore More

- [Raw Qwen3 4B Instruct 2507 direct logits profile](/models/jevbench-raw-qwen3-4b-instruct-2507)
- [typecastlm (Mikhail Gribov, Qwen3.5-4B computed head) profile](/models/jevbench-typecastlm)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)